Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce...

105
Outline Scheduling problem description State-of-art approaches Further scheduling topics Scheduling and Information Systems Multi-skilled workforce management Murat Fırat PhD., Discrete Mathematics Group, TU/e Post-doc, France T´ el´ ecom Orange Labs., July 25,2017 Seminar, Information Systems Group, TU/e Murat Fırat Multi-skilled workforce management

Transcript of Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce...

Page 1: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Multi-skilled workforce management

Murat Fırat

PhD., Discrete Mathematics Group, TU/e

Post-doc, France Telecom Orange Labs.,

July 25,2017

Seminar, Information Systems Group, TU/e

Murat Fırat Multi-skilled workforce management

Page 2: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling problem descriptionMotivationBasic conceptsComplexity

State-of-art approachesALNS approachLocal search approachFMM approachComputational Results

Further scheduling topicsStability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Scheduling and Information SystemsScheduling and Business Process AnalysisScheduling and Artificial Intelligence

Murat Fırat Multi-skilled workforce management

Page 3: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Motivation for advanced scheduling

Murat Fırat Multi-skilled workforce management

Page 4: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitiveI to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 5: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitiveI to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 6: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitiveI to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 7: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitiveI to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 8: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitive

I to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 9: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Advanced scheduling in France Telecom

I Steadily increasing number of services.

I Employing more than 105 technicians.

I End of the telecommunication monopoly in France.

I So, the emerging need is

I to be more competitiveI to limit the growth of the technician group

Murat Fırat Multi-skilled workforce management

Page 10: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Basic concepts in scheduling data

Murat Fırat Multi-skilled workforce management

Page 11: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks with

I Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians with

I AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 12: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing options

I Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians with

I AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 13: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians with

I AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 14: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians with

I AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 15: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI Availabilities

I Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 16: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 17: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI AvailabilitiesI Skills with hierarchical levels

I Our mission

I Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 18: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI AvailabilitiesI Skills with hierarchical levels

I Our missionI Outsource some tasks

I Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 19: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI AvailabilitiesI Skills with hierarchical levels

I Our missionI Outsource some tasksI Schedule all tasks on a day-to-day basis

I Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 20: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

I Given a set J = {j1, j2, . . . } of tasks withI Precedence relations, priority classes, outsourcing optionsI Skill requirements

I Also, a set T = {t1, t2, . . . } of technicians withI AvailabilitiesI Skills with hierarchical levels

I Our missionI Outsource some tasksI Schedule all tasks on a day-to-day basisI Minimize the weighted makespan

Murat Fırat Multi-skilled workforce management

Page 21: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

More on Skills:

I Skill categories/domains: D = {d1, d2, . . . }.

I Hierarchical skill levels: L = {l0, l1, . . . }.

I Skills of technician t are denoted by Skt ∈ {0, 1}|L|×|D|

I Skill requirements of task j are denoted by Rqj ∈ Z|L|×|D|

Murat Fırat Multi-skilled workforce management

Page 22: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

More on Skills:

I Skill categories/domains: D = {d1, d2, . . . }.

I Hierarchical skill levels: L = {l0, l1, . . . }.

I Skills of technician t are denoted by Skt ∈ {0, 1}|L|×|D|

I Skill requirements of task j are denoted by Rqj ∈ Z|L|×|D|

Murat Fırat Multi-skilled workforce management

Page 23: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

More on Skills:

I Skill categories/domains: D = {d1, d2, . . . }.

I Hierarchical skill levels: L = {l0, l1, . . . }.

I Skills of technician t are denoted by Skt ∈ {0, 1}|L|×|D|

I Skill requirements of task j are denoted by Rqj ∈ Z|L|×|D|

Murat Fırat Multi-skilled workforce management

Page 24: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

More on Skills:

I Skill categories/domains: D = {d1, d2, . . . }.

I Hierarchical skill levels: L = {l0, l1, . . . }.

I Skills of technician t are denoted by Skt ∈ {0, 1}|L|×|D|

I Skill requirements of task j are denoted by Rqj ∈ Z|L|×|D|

Murat Fırat Multi-skilled workforce management

Page 25: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

More on Skills:

I Skill categories/domains: D = {d1, d2, . . . }.

I Hierarchical skill levels: L = {l0, l1, . . . }.

I Skills of technician t are denoted by Skt ∈ {0, 1}|L|×|D|

I Skill requirements of task j are denoted by Rqj ∈ Z|L|×|D|

Murat Fırat Multi-skilled workforce management

Page 26: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

Consider |D| = |L| = 3 such that

I Technicians t1 and t2 with skills

Skt1 =

d1 d2 d3

l0 1 1 1l1 1 0 0l2 1 0 0

,Skt2 =

d1 d2 d3

l0 1 1 1l1 0 1 1l2 0 1 0

I Tasks j1 and j2 with skill requirements

Rqj1 =

d1 d2 d3

l0 1 2 2l1 1 1 1l2 1 0 0

,Rqj2 =

d1 d2 d3

l0 1 1 1l1 1 1 0l2 1 1 0

I So, use the team τ = {t1, t2}!

Murat Fırat Multi-skilled workforce management

Page 27: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

Consider |D| = |L| = 3 such that

I Technicians t1 and t2 with skills

Skt1 =

d1 d2 d3

l0 1 1 1l1 1 0 0l2 1 0 0

,Skt2 =

d1 d2 d3

l0 1 1 1l1 0 1 1l2 0 1 0

I Tasks j1 and j2 with skill requirements

Rqj1 =

d1 d2 d3

l0 1 2 2l1 1 1 1l2 1 0 0

,Rqj2 =

d1 d2 d3

l0 1 1 1l1 1 1 0l2 1 1 0

I So, use the team τ = {t1, t2}!

Murat Fırat Multi-skilled workforce management

Page 28: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

Consider |D| = |L| = 3 such that

I Technicians t1 and t2 with skills

Skt1 =

d1 d2 d3

l0 1 1 1l1 1 0 0l2 1 0 0

,Skt2 =

d1 d2 d3

l0 1 1 1l1 0 1 1l2 0 1 0

I Tasks j1 and j2 with skill requirements

Rqj1 =

d1 d2 d3

l0 1 2 2l1 1 1 1l2 1 0 0

,Rqj2 =

d1 d2 d3

l0 1 1 1l1 1 1 0l2 1 1 0

I So, use the team τ = {t1, t2}!

Murat Fırat Multi-skilled workforce management

Page 29: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Scheduling data of France Telecom

Consider |D| = |L| = 3 such that

I Technicians t1 and t2 with skills

Skt1 =

d1 d2 d3

l0 1 1 1l1 1 0 0l2 1 0 0

,Skt2 =

d1 d2 d3

l0 1 1 1l1 0 1 1l2 0 1 0

I Tasks j1 and j2 with skill requirements

Rqj1 =

d1 d2 d3

l0 1 2 2l1 1 1 1l2 1 0 0

,Rqj2 =

d1 d2 d3

l0 1 1 1l1 1 1 0l2 1 1 0

I So, use the team τ = {t1, t2}!

Murat Fırat Multi-skilled workforce management

Page 30: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

A workday schedule

Murat Fırat Multi-skilled workforce management

Page 31: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

A complete schedule

The schedule cost is the weighted makespan:∑

i wiCi .

Murat Fırat Multi-skilled workforce management

Page 32: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

How hard is to solve our schedulingproblem?

Murat Fırat Multi-skilled workforce management

Page 33: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Theoretical result

Theorem 1Technician scheduling problem of France Telecom isNP-Hard in the strong sense.1

1Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

Murat Fırat Multi-skilled workforce management

Page 34: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

Benchmark instances2

Data set A Data set B Data set X

Ins. |T | |J| |D| |L| |T | |J| |D| |L| |T | |J| |D| |L|1 5 5 3 2 20 200 4 4 60 600 15 42 5 5 3 2 30 300 5 3 100 800 6 63 7 20 3 2 40 400 4 4 50 300 20 34 7 20 4 3 30 400 40 3 70 800 15 75 10 50 3 2 50 500 7 4 60 600 15 46 10 50 5 4 30 500 8 3 20 200 6 67 20 100 5 4 100 500 10 5 50 300 20 38 20 100 5 4 150 800 10 4 30 100 15 79 20 100 5 4 60 120 5 5 50 500 15 4

10 15 100 5 4 40 120 5 5 40 500 15 4

2Technicians and interventions scheduling for telecommunications, FranceTelecom R&D, Orange Labs.

Murat Fırat Multi-skilled workforce management

Page 35: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

MotivationBasic conceptsComplexity

What about formulating as a MILP model?

An experimentation3reports

After 24-hour computation time, instances A3 and A4 couldnot be solved by CPLEX 11, leaving optimality gaps 20% and15% respectively.

3Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 36: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

State-of-art approaches to ourscheduling problem

Murat Fırat Multi-skilled workforce management

Page 37: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:

I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:

I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 38: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..

I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:

I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 39: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:

I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 40: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:

I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 41: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:I Choose a destroy and a repair method.

I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 42: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.

I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 43: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 44: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Adapted Large Neighborhood search 4

I A construction heuristic for initial schedule:I Construct teams for seed tasks..I Criteria for seed tasks: Criticality, difficulty, similarity.

I Modifying the current schedule:I Choose a destroy and a repair method.I Accepting a new schedule: Use simulated annealing criterion.I Update scores of destroy and repair methods.

I Within timelimit: Make restarts of the schedule modification.

4Scheduling technicians and tasks in a telecommunication company,Cordeau, J. F., Laporte, G., Pasin F., Ropke, S., 2010, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 45: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithm

I Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.I Swapping tasks randomly within a day, between days.I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 46: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithmI Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.I Swapping tasks randomly within a day, between days.I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 47: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithmI Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.

I Swapping tasks randomly within a day, between days.I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 48: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithmI Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.I Swapping tasks randomly within a day, between days.

I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 49: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithmI Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.I Swapping tasks randomly within a day, between days.I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 50: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

High-Performance local search heuristic5

I Obtain an initial schedule using a greedy algorithmI Modify the schedule with 31 predefined moves:

I Swapping technicians randomly.I Swapping tasks randomly within a day, between days.I Other 28 sophisticated moves.

I Some bookkeeping for maintaining precedence relations.

5 High-Performance local search for task scheduling with human resourceallocation, Estellon, B., Gardi, F., Nouioua, K., 2009, Lecture Notes InComputer Science.

Murat Fırat Multi-skilled workforce management

Page 51: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 52: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 53: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.

I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 54: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 55: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloads

I merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 56: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloads

I reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 57: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a team

I initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 58: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 59: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM) Approach6

General properties:

I Constructing day schedules one by one.

I Initializing a day schedule with one-task team loads.I Inserting more tasks into day schedule possibly by

I extending teamloadsI merging teamloadsI reshuffling technicians of a teamI initializing a new team

I The above decisions are simultaneously made by a flexibleMatching model!

6An improved MIP-based approach for a multi-skill workforce schedulingproblem, Fırat., M., Hurkens, C., 2012, Journal of Scheduling.

Murat Fırat Multi-skilled workforce management

Page 60: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Initial Matching Model (IMM)

Murat Fırat Multi-skilled workforce management

Page 61: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

IMM: Bipartite Graph

Murat Fırat Multi-skilled workforce management

Page 62: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

IMM: Matching constraints

Murat Fırat Multi-skilled workforce management

Page 63: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

IMM: Matching solution

Murat Fırat Multi-skilled workforce management

Page 64: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

IMM: Initialized teams

Murat Fırat Multi-skilled workforce management

Page 65: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM Approach

Matching weights of tasks are sum of the following criteria

I processing time

I coverage

I min-tech

I hardness

Murat Fırat Multi-skilled workforce management

Page 66: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM Approach

Matching weights of tasks are sum of the following criteria

I processing time

I coverage

I min-tech

I hardness

Murat Fırat Multi-skilled workforce management

Page 67: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM Approach

Matching weights of tasks are sum of the following criteria

I processing time

I coverage

I min-tech

I hardness

Murat Fırat Multi-skilled workforce management

Page 68: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM Approach

Matching weights of tasks are sum of the following criteria

I processing time

I coverage

I min-tech

I hardness

Murat Fırat Multi-skilled workforce management

Page 69: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM Approach

Matching weights of tasks are sum of the following criteria

I processing time

I coverage

I min-tech

I hardness

Murat Fırat Multi-skilled workforce management

Page 70: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Flexible Matching Model (FMM)

Murat Fırat Multi-skilled workforce management

Page 71: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM: Bipartite Graph

Murat Fırat Multi-skilled workforce management

Page 72: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM: Matching skills or loads of teams

Murat Fırat Multi-skilled workforce management

Page 73: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM: Matching solution

Murat Fırat Multi-skilled workforce management

Page 74: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

FMM: Extended day schedule

Murat Fırat Multi-skilled workforce management

Page 75: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Computational Results: Set A

Murat Fırat Multi-skilled workforce management

Page 76: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Computational Results: Set B

Murat Fırat Multi-skilled workforce management

Page 77: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

ALNS approachLocal search approachFMM approachComputational Results

Computational Results: Set X

Murat Fırat Multi-skilled workforce management

Page 78: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Further scheduling topics

Murat Fırat Multi-skilled workforce management

Page 79: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Stability in multi-skilled workforce assignments

I The notion of stability in workforce assignments is defined7.

I Special case analysis is conducted.

I Problem complexity is established.

I The problem is formulated as an IP model.

I A Branch-and-Price algorithm is developed8.

7Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

8A Branch-and-Price algorithm for stable multi-skill workforce assignmentswith hierarchical skills, Fırat, M., Briskorn, D., Laugier, A., 2016, EJOR

Murat Fırat Multi-skilled workforce management

Page 80: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Stability in multi-skilled workforce assignments

I The notion of stability in workforce assignments is defined7.

I Special case analysis is conducted.

I Problem complexity is established.

I The problem is formulated as an IP model.

I A Branch-and-Price algorithm is developed8.

7Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

8A Branch-and-Price algorithm for stable multi-skill workforce assignmentswith hierarchical skills, Fırat, M., Briskorn, D., Laugier, A., 2016, EJOR

Murat Fırat Multi-skilled workforce management

Page 81: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Stability in multi-skilled workforce assignments

I The notion of stability in workforce assignments is defined7.

I Special case analysis is conducted.

I Problem complexity is established.

I The problem is formulated as an IP model.

I A Branch-and-Price algorithm is developed8.

7Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

8A Branch-and-Price algorithm for stable multi-skill workforce assignmentswith hierarchical skills, Fırat, M., Briskorn, D., Laugier, A., 2016, EJOR

Murat Fırat Multi-skilled workforce management

Page 82: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Stability in multi-skilled workforce assignments

I The notion of stability in workforce assignments is defined7.

I Special case analysis is conducted.

I Problem complexity is established.

I The problem is formulated as an IP model.

I A Branch-and-Price algorithm is developed8.

7Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

8A Branch-and-Price algorithm for stable multi-skill workforce assignmentswith hierarchical skills, Fırat, M., Briskorn, D., Laugier, A., 2016, EJOR

Murat Fırat Multi-skilled workforce management

Page 83: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Stability in multi-skilled workforce assignments

I The notion of stability in workforce assignments is defined7.

I Special case analysis is conducted.

I Problem complexity is established.

I The problem is formulated as an IP model.

I A Branch-and-Price algorithm is developed8.

7Stable multi-skill workforce assignments, Fırat, M., Hurkens, C., Laugier,A., 2014, Annals of OR.

8A Branch-and-Price algorithm for stable multi-skill workforce assignmentswith hierarchical skills, Fırat, M., Briskorn, D., Laugier, A., 2016, EJOR

Murat Fırat Multi-skilled workforce management

Page 84: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Scheduling multi-skilled workforce with varyingperformances9

I Scheduling by taking a dynamic view of human performance.

I It is shown how the scheduling problem can be constructedfrom business process knowledge.

I The types of nodes in process trees matched with the notionof scheduling concepts.

I The problem is formulated as an MIP model to conductcomputational experimentation.

9Human Performance-Aware Scheduling and Routing of a Multi-SkilledWorkforce, van Eck, M., Fırat, M., Nuijten, W., Sidorova, N., van der Aalst,W. , submitted to CSIMQ.

Murat Fırat Multi-skilled workforce management

Page 85: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Scheduling multi-skilled workforce with varyingperformances9

I Scheduling by taking a dynamic view of human performance.

I It is shown how the scheduling problem can be constructedfrom business process knowledge.

I The types of nodes in process trees matched with the notionof scheduling concepts.

I The problem is formulated as an MIP model to conductcomputational experimentation.

9Human Performance-Aware Scheduling and Routing of a Multi-SkilledWorkforce, van Eck, M., Fırat, M., Nuijten, W., Sidorova, N., van der Aalst,W. , submitted to CSIMQ.

Murat Fırat Multi-skilled workforce management

Page 86: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Scheduling multi-skilled workforce with varyingperformances9

I Scheduling by taking a dynamic view of human performance.

I It is shown how the scheduling problem can be constructedfrom business process knowledge.

I The types of nodes in process trees matched with the notionof scheduling concepts.

I The problem is formulated as an MIP model to conductcomputational experimentation.

9Human Performance-Aware Scheduling and Routing of a Multi-SkilledWorkforce, van Eck, M., Fırat, M., Nuijten, W., Sidorova, N., van der Aalst,W. , submitted to CSIMQ.

Murat Fırat Multi-skilled workforce management

Page 87: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Scheduling multi-skilled workforce with varyingperformances9

I Scheduling by taking a dynamic view of human performance.

I It is shown how the scheduling problem can be constructedfrom business process knowledge.

I The types of nodes in process trees matched with the notionof scheduling concepts.

I The problem is formulated as an MIP model to conductcomputational experimentation.

9Human Performance-Aware Scheduling and Routing of a Multi-SkilledWorkforce, van Eck, M., Fırat, M., Nuijten, W., Sidorova, N., van der Aalst,W. , submitted to CSIMQ.

Murat Fırat Multi-skilled workforce management

Page 88: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Pilot workforce planning of an airline company in Turkey

I An aggregated planning problem of pilot in a one-year horizon.

I Demands for every aircraft type are to be met.

I Dynamic skills of pilots with trainings.

I Advise a pilot employment plan to human resourcedepartment.

Murat Fırat Multi-skilled workforce management

Page 89: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Pilot workforce planning of an airline company in Turkey

I An aggregated planning problem of pilot in a one-year horizon.

I Demands for every aircraft type are to be met.

I Dynamic skills of pilots with trainings.

I Advise a pilot employment plan to human resourcedepartment.

Murat Fırat Multi-skilled workforce management

Page 90: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Pilot workforce planning of an airline company in Turkey

I An aggregated planning problem of pilot in a one-year horizon.

I Demands for every aircraft type are to be met.

I Dynamic skills of pilots with trainings.

I Advise a pilot employment plan to human resourcedepartment.

Murat Fırat Multi-skilled workforce management

Page 91: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Stability in multi-skilled workforce assignmentsScheduling multi-skilled workforce with varying performancesPilot workforce planning of an airline company in Turkey

Pilot workforce planning of an airline company in Turkey

I An aggregated planning problem of pilot in a one-year horizon.

I Demands for every aircraft type are to be met.

I Dynamic skills of pilots with trainings.

I Advise a pilot employment plan to human resourcedepartment.

Murat Fırat Multi-skilled workforce management

Page 92: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Information Systems

Murat Fırat Multi-skilled workforce management

Page 93: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Precision in scheduling data

Setup times

I are the components of scheduling data due to real-life issue.

I usually obtained by rough approximations.

I can be extracted using some historical data of companies.

I possibly one topic that may integrate Scheduling and BusinessProcess Analysis

Murat Fırat Multi-skilled workforce management

Page 94: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Precision in scheduling data

Setup times

I are the components of scheduling data due to real-life issue.

I usually obtained by rough approximations.

I can be extracted using some historical data of companies.

I possibly one topic that may integrate Scheduling and BusinessProcess Analysis

Murat Fırat Multi-skilled workforce management

Page 95: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Precision in scheduling data

Setup times

I are the components of scheduling data due to real-life issue.

I usually obtained by rough approximations.

I can be extracted using some historical data of companies.

I possibly one topic that may integrate Scheduling and BusinessProcess Analysis

Murat Fırat Multi-skilled workforce management

Page 96: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Precision in scheduling data

Setup times

I are the components of scheduling data due to real-life issue.

I usually obtained by rough approximations.

I can be extracted using some historical data of companies.

I possibly one topic that may integrate Scheduling and BusinessProcess Analysis

Murat Fırat Multi-skilled workforce management

Page 97: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Precision in scheduling data

Setup times

I are the components of scheduling data due to real-life issue.

I usually obtained by rough approximations.

I can be extracted using some historical data of companies.

I possibly one topic that may integrate Scheduling and BusinessProcess Analysis

Murat Fırat Multi-skilled workforce management

Page 98: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 99: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 100: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 101: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 102: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 103: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 104: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Scheduling and Artificial Intelligence

Exact algorithms

I are smart enumerations methods

I provide us optimality gaps

I usually decomposes scheduling problem into sub-problems

I there are more than one criteria to deal with

I the sub-problems are usually finding “good” objects

I defining adaptively how an object is “good” is crucial

Murat Fırat Multi-skilled workforce management

Page 105: Multi-skilled workforce managementmfirat/Presentations/MFirat_MWM_25_07...Multi-skilled workforce management Murat F rat PhD., Discrete Mathematics Group,TU/e Post-doc, France T el

OutlineScheduling problem description

State-of-art approachesFurther scheduling topics

Scheduling and Information Systems

Scheduling and Business Process AnalysisScheduling and Artificial Intelligence

Thanks

Murat Fırat Multi-skilled workforce management