Stochastic Programming and Scenario Generation: Decision Modelling Simulation and Information systems Perspective

Stochastic Programming brings together models of optimum resource allocation and models of randomness and thereby creates a robust decision making framework. The models of randomness with their finite, discrete realisations are known as Scenario Generators. In this thesis we consider alternative approaches to Scenario Generation (SG) in a generic form which can be used to formulate (a) two stage (static) and (b) multistage dynamic SP models. We also investigate the modelling structure and software issues of integrating a scenario generator with an optimisation model to construct Stochastic Programming recourse problems. We consider how the Expected Value (EV) and Stochastic Programming (SP) decision model results can be evaluated within a descriptive modelling framework of simulation. Illustrative examples and computational results are given in support of our investigation. We set out a blueprint for a SG model library; its open architecture and how it is used in decision models and in simulation.
The major contribution of this research reported in this thesis is the development of a two phase modelling paradigm which connects forward looking (ex ante) decision making with wait and see (ex post) results analysis. The research also connects these modelling paradigms to the analytic information systems, which are used by decision makers of diverse organisations. This latter aspect, which integrates information engineering with analytics decision models taking into consideration uncertainty and risk, is a contribution to knowledge. The scenario generation, the simulation extension to SAMPL and the design of the SG library encapsulate a number of novel features and are also a contribution to knowledge.