Lecture 4-INSE6400 Flashcards
Decision Making Modeling and Analysis
Decision Making Process
Planning Decision Making Process >
Gathering Data > Organizing and Processing Info > Making Decision >
Implementing Decision
What are the factors involve in Decision Making Process?
- Goals and objectives:
Provide the basis for comparing different choices.
Provide information for trade-off.
Decision should made to satisfy the importantstakeholders. - Decision type:
Binary decision (go or no-go) vs. Selection among somechoices (make or buy and supplier selection)
Who make the decision: individual or group?
- Decision context:
Scope of the decision
Context dimensions: technical, financial, personnel,process, temporal and legacy - Stakeholders:
Anyone (people or organization) who will beaffected by the results of the decision.
Important aspect in policy making. - Legacy decisions:
Learn from the past. - Supporting data:
Ensure proper information to support the decision.
Data collection may require a lot of resources.
Accuracy in data collected depends on the decisiontype and context.
Decision Framework: Type of Decision.
3 Types: Structured, Semi-structured and Unstructured.
Structured decision making:
Routine, with well-understood context and scope. E.g., selecting a meal.
Semi-structured decision making:
Past decisions cannot be repeatedly used. E.g., buying a car
Unstructured decision making:
Complex problems that are unique and typical one-time. E.g., adopting a new technology, developing iPod ornot.
Scope of Control in Decision Making:
3 Scopes. Operational, Managerial and Strategic.
Operational: Practitioner level on some daily or routine work.
Managerial: Define the management, mentoring or coaching level ofdecisions. E.g., when formulating a project, decide on the budgetand time.
Strategic: Represent an executive or enterprise level control. E.g., investment decisions, merging of company,whether or not to develop a new product.
Decisions Making Models:
Definition of model: a physical, mathematical, or otherwise logicalrepresentationof a system entity, phenomenon orprocess.
Physical, Mathematical or Other.
Mathematical models:
Use mathematical notation to represent a relationship orfunction.
Computable: quantitative and use of algorithms
Physical models:
Reflect directly some or most of the physical characteristicsof the actual system or system element under study.
Prototype: can be both hardware and software.
Representation model = abstraction, only approximation.
Schematic models.
Diagrams or charts representing a system element or process.
E.g., block diagram, context diagram, functional flow block diagram.
Physical vs Mathematical model Examples:
Mathematical models:
Differential equations for dynamic systems (finiteelement analysis, computational fluid dynamics,population model)
Discrete event simulation (traffic simulation)
Formal language in computer science.
Physical model:
Scale model of airplane for the simulation in thewind tunnel
Vehicle used in crash tests.
3D printing.
Schematic Models Characteristics (They are other type of Models)
Common examples are Traditional Hierarchy Block Diagram, Context Diagram, Functional Flow Block Diagram)
Intuitive, you don’t need in-depth training forunderstanding.
Good for conceptual clarityand understand.
Not good to supportcomputing and analysis.
Simulation Model:
Simulation Definition: experiments on the (computational)models.
Key benefit: understand and predict the systembehavior or propertywithout the cost of buildingphysical prototypes.
Types of Simulation Models:
- Operational simulation:
Simulation of operational systems. used in conceptualdevelopment stage. - Games Simulation of the interactions between (competing)entities:
Game with the environment (or nature). Examine and decide thegame strategyfor betteroutcomes.
- Physical simulation: Study the physical behavior of system elementsE.g.,vehicle motion with external forces.
- Hardware-in-the-loop simulation:
Physical simulation in whichactual system hardwareiscoupled with a computer-driven simulation. - Environmental simulation:
Used in engineering test and evaluation.
Mechanical (computational) stress testing, crash testing,wind tunnel testing. - Virtual reality simulation:
Use of 3D visual environment
E.g., pilot simulator.
Simulation Key Facts:
Development of system simulations: Simulation is a complex subject.
Balance betweenfidelityandcomplexity.
Design of experimentsto control the experimentalsetups.
Simulation verification and validation:
Verification: check the consistency within thesystem’s domain.
Validation: check whether the simulation can fairlycorrespond to the real world (environment).
Trade-off Analysis: Basic Trade-off Analysis: (Fuel-efficient car vs. powerful car)
Defining the objectives:Expressed as quantitative measures.
At leasttwo objectives for a possible trade-off.
Identification of alternatives:
Include all promising candidate alternatives.
Comparing the alternatives
Determine therelative meritsof the alternatives: Seek for a balance solution.
Sensitivity analysis:
Do not just select the “top” winner. Any alternatives thatperform close to the winner should be re-examinedcarefully.
Formal Trade-off Analysis:
Step 1: Definition of the objectives
Step 2: Identification of viable alternatives: There is never a single possible solution.
Optimal solution vs. “good-enough” solution.
Understand thediscriminatorsamong alternatives.
Remain opento addition solution surfacing during the tradestudy.
Step 3: Definition of selection criteria
Quantifiable and objective measure is one key. E.g., cost, reliability, maintainability, ease of use, etc.
Step 4: Assignment ofweighting factorsto selectioncriteria
Weighting factors( different importance of criteria).
E.g., 5-point scale, limit the sum of weights, normalization.
Step 5: Assignment ofvalue ratingsforalternatives:
Why: each criterion may use different units.
Three options: subjective value method, step functionmethod and utility function method.
Step 6: Calculating comparative scores:
Basic method: weighted sum
Using “just” a single sum to select an alternative may beover-simplified.
Step 7: Analyzing the results:
Why: qualitative judgments and incommensurable natureof criteria
Sensitivity analysis.
Examples of Quantification:
Reliability:
Component A has 95% chance being functional for 5years.
This data can be obtained from past records andexperiments.
Maintainability:
The system is running properly 97% of the service time.
MTBF = mean time between failures
MTTR = mean time to repair.
Ease of use:
Number of steps (or time) required to complete sometasks.
Assignment of Value Ratings:
Three options: subjective value method, step functionmethod and utility function method.
Subjective valuemethod:
5-point scale 1= poor 2 = fair 3 = satisfactory 4=good 5= superior.
Actual measurementmethod:
An explicit mappingbetween a measure and ascore.
Utility function method:
Mapped to the scorebetween 0 and 1.
Normalization takes place.
Math Based on Assignment of Value Ratings