Decision Support Systems Management Information Systems BUS 391 Barry Floyd.
-
Upload
blake-armstrong -
Category
Documents
-
view
215 -
download
0
Transcript of Decision Support Systems Management Information Systems BUS 391 Barry Floyd.
Decision Support Systems
Management Information SystemsBUS 391Barry Floyd
Agenda
Excel ExamplesDefinitionFundamentalsConclusion
Tax Computation
Given certain assumptions, earnings, and savings goals, how much should John and Sue pay in estimated quarterly taxes?
Tax Computation
Save 5% of total income in a tax deductible retirement account, up to a maximum of $3,000Entitled to personal exemption of $3,100 eachStandard deduction for joint tax filers is $8,000Tax brackets 15% for up to $48,000 and 26% for $48,001 to $115,000How much estimated taxes should they pay each quarter?
Parameters
Problem ParametersRetirement Saviings 0.05Maximum savings 3000Personal Examption per person 3100Standard Deduction 8000Tax brackets 0.15 up to 48000
0.26 up to 115000
Input area
Input DataDavid's Income 45000Sue's Income 40000
Tax Computation
Tax ComputationTotal Income 85000Retirement Savings 3000Personal Exemptions 6200Standard Deduction 8000Taxable Income 67800Tax @ 15% rate 7200Tax @ 26% rate 5148Total Tax 12348Estimated Tax per quarter 3087
Breakeven Analysis
Determine Total RevenueFixed CostTotal Variable CostTotal CostProfit
Known Parameters
Know ParametersSelling price per unit $10.00Fixed cost $1,000.00Variable cost per unit $5.00
Input
Input DataNumber of units, X 2200
Results
ResultsTotal revenue $22,000.00Fixed cost $1,000.00Total variable cost $11,000.00total cost $12,000.00Profit $10,000.00
What is a DSS?
An interactive information system that provides information, models, and data manipulation tools to help make decisions in semi-structured and unstructured situations where no one know exactly how the decision should be made.
Steps involved in Decision Modeling
Formulation Defining the problem Developing a model Acquiring Input data
Solution Developing a solution Testing the solution
Interpretation Analyzing the results and sensitivity analysis Implementing the results
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Develop a clear, concise statement of
the problem. Go beyond symptoms,
look for cause! Find measurable objectives.
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Develop a model … for decisions
modeling, this is a mathematical
model.Decision variable is
controllable, a parameter is an
inherent measurable quantity.
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Get data from reports or interviews
or sampling, etc. (e.g., time to
manufacture a widget).
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Manipulate model to arrive at the best (or optimal) solution to
the problem.
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Test completely. Use known data, comparison data,
etc.
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Determine implications of solution. What
happens if results are implemented?
How sensitive is the solution to
fluctuations?
Steps involved in Decision Modeling
• Formulation– Defining the problem– Developing a model– Acquiring Input data
• Solution– Developing a solution– Testing the solution
• Interpretation– Analyzing the results and sensitivity analysis– Implementing the results
Can be the most difficult part …
Different flavors of DSS
Simulation and optimizationOLAP and Data MiningExpert SystemsNeural NetworksFuzzy LogicCase-based ReasoningIntelligent Agents
Simulation and optimization
Simulation – calculates outcomes based on some abstraction (typically mathematical) of the situationOptimization – calculates the ‘best’ answer given certain sets of constraints (e.g., which set of fixed and variable costs given a range of potential sales would provide the most profit).
OLAP and Data Mining
OLAP – Online analytical processing Explores large volumes of transaction data
Data Mining Explores large volumes of data looking for
patterns that help managers understand critical relationships Eg if someone buys cake mix, do they also buy
frosting? What drives paint sales? Not new home
purchases, but sales of existing homes.
Expert Systems
Builds on ‘knowledge’ typically extracted from an expert (e.g., a medical specialist on cancer) Knowledge must be ‘captured’ and
represented within the system Typically done with If-Then rules
Data about particular case Inference engine applies rules to data
to derive an outcome
Neural Networks
Statistical method for finding and representing patterns in dataNeural represents the way researchers believe the brain works A neural network is an information
system that recognizes objects or patterns based on examples that have been used to train it.
Fuzzy Logic
Fuzzy logic is a form of reasoning that makes it possible to combine imprecise conditions stated in a form similar to the types of descriptive categories people use. Don’t use either/or logic, allow categories to
be somewhat vague and potentially overlapping: Very profitable, somewhat profitable, slightly profitable categories.
Use multiple rules and build a system that combines the rules in a meaningful manner.
Case Based Reasoning
A DSS method based on the idea of finding past cases most similar to the current situation in which a decision must be made. Must maintain a Database of cases and
a means of searching the cases to match the problem at hand.
Must have a means of ‘categorizing’ cases and limiting structure to a manageable set.
Intelligent Agents
An autonomous, goal-directed computerized process that can be launched into a computer system/network to do background work Shopbots, email agents, news agents.
Shopbot – find best price for X Email – scan email messages as they arrive and
determine if user should be interrupted News Agent – scan news sources to put together
a customized newspaper.
Conclusion
DSS is decided different than TPS and MISWe’ll employ Excel as our modeling/DSS environment
Data Tableand Scenario Management
Barry Floyd
Data Tables and Scenario Management
Data Table Displays results of multiple what-if analyses
One variable Data Table Specify one input cell and any number of result cells
Two variable Data Table Specify two input cells and one result cell
Scenario Allows you to define a set of input cells and
result cells and to then view the results in a systematic fashion
One Variable Data Table
Number of units Total Revenue Total Cost Profit2200 $22,000.00 $14,200.00 $7,800.002300 $23,000 $14,800 $8,2002400 $24,000 $15,400 $8,6002500 $25,000 $16,000 $9,0002600 $26,000 $16,600 $9,400
One or more result cellsOne input
cell
One Variable Data Table
Steps Create row of output formulas Define column of input values Highlight formulas and input values Select data, table Indicate the 'input' cell - note we have a column of
values and so choose column
Two Variable Tables
-Number of Units 10 10.5 11 11.5 12 12.5 132300 $10,500 $11,650 $12,800 $13,950 $15,100 $16,250 $17,4002400 $11,000 $12,200 $13,400 $14,600 $15,800 $17,000 $18,2002500 $11,500 $12,750 $14,000 $15,250 $16,500 $17,750 $19,0002600 $12,000 $13,300 $14,600 $15,900 $17,200 $18,500 $19,8002700 $12,500 $13,850 $15,200 $16,550 $17,900 $19,250 $20,600
Selling Price per Unit
Variable 1 values in first row
Variable 2 values in
first Column
Result values
displayed in table
Two variable tables
Steps in creating a two variable data table Create column of variable costs Create row of number of units Place "RESULT CELL" reference in the upper left hand
corner Format cell to show a label rather than the formula Add a label Highlight table area Select data, select table Assign B4 to row input cell, B9 to column input cell Format table to currency
Scenario Manager
Used to perform ‘what-if’ analyses given more than two variablesIdentify key variables whose values are important for characterizing the ‘scene’ High quality
Fixed costs are higher, variable costs are higher, Selling price is higher
Low quality Fixed costs lower, variable costs are medium, selling
price is low Mid quality
Fixed costs mid, variable costs are medium, selling price is medium
Values
High Medium Low
Selling Price
$50 $30 $10
Fixed Costs
$5000 $2500 $1000
Variable Costs
$20 $10 $3
Steps
Select tools, scenariosCreate a scenarioAdd values to attributesRepeat for each scenarioClick on a scenario, click showOr click on summary
Output results
Scenario SummaryCurrent Values: High Quality Medium Quality Low Quality
Changing Cells:Variable_cost_per_unit $5.00 $20.00 $10.00 $3.00Selling_price_per_unit $10.00 $50.00 $30.00 $10.00$B$5 $1,000.00 $5,000.00 $2,500.00 $1,000.00
Result Cells:Profit $10,000.00 $61,000.00 $41,500.00 $14,400.00
Summary
Use the power of excel to analyze data in an interactive format.Do ‘what-if’ analyses on a one variable, two variable and multi-variable format.Very powerful, relatively easy to use.