Distance MSc in Operational Research - Decision Analysis

Credit Value: 6
Tutors: Tim Bedford

Class description/introduction

The basic ideas of decision making under uncertainty are covered. Graphical modelling using Bayesian Nets (Influence diagrams and Belief nets) are used to structure models. The basic notions of probability, utility and value of information are covered.

Class aims

To enable students to build models that support decision making under uncertainty, using decision analysis software and taking account of different stakeholder perceptions.

Learning outcomes

Subject specific knowledge and skills

• Learn how to structure a model with decisions, uncertainties and consequences
• Build graphical models of decision problems and exploit them to solve problems, including the use of sensitivity analysis
• Understand and apply Expected Monetary Value and Expected Utility criteria to support a decision maker
• Apply value of information techniques to support a decision maker in spending resources to reduce uncertainties
• Use of different criteria and sensitivity analysis to explore the impact of different stakeholder viewpoints

Cognitive abilities and non-subject specific skills

• Improved problem structuring skills
• Report writing

Content/Structure of class

Various different decision criteria are introduced, and the ideas of decision trees, leading into discussion of formal approaches to the analysis of problems involving uncertainty and multiple criteria. The modelling of decision problems with uncertain consequences using influence diagrams is discussed. In the context of this course, influence diagrams are used to represent probabilistic influence. The software is used to complete an assignment.

Reading List

The following may be of use but are not required:

• Simon French "Decision Theory: An Introduction to the Mathematics of Rationality", Ellis Horwood, Chichester, 1988.
• Robert Clemen "Making Hard Decisions", Duxbury, Wadsworth, 1991.
• Uncertainty; a guide to dealing with uncertainty in quantitative risk and policy analysis Morgan, M.Granger. and Henrion, Max, Cambridge, UK : Cambridge University Press, 1990.
• DPL manual

Assessment

Assignment made using the software DPL, consisting of a written client report supported by computer analysis.

Learning Outcomes

• Learn how to structure a model with decisions, uncertainties and consequences
• Build graphical models of decision problems and exploit them to solve problems, including the use of sensitivity analysis
• Understand and apply Expected Monetary Value and Expected Utility criteria to support a decision maker
• Apply value of information techniques to support a decision maker in spending resources to reduce uncertainties
• Improved problem structuring skills
• Report writing

management science DEPARTMENT OF MANAGEMENT SCIENCE
UNIVERSITY OF STRATHCLYDE Graham Hills BUILDING 40 GEORGE STREET G1 1QE
t:0141 548 3613/3141 f:0141 552 6686
contact-mansci@strath.ac.uk