Computer based Approaches to Managerial Decision making:optimism, heuristics and simulation

Abstract

This research studies three computer based approaches - optimisation, heuristic, and simulation, to numerical optimisation problem in managerial decision making. In Chapter 2 we first discuss computer based mathematical programming with the focus on linear programming (LP) and its software in the application of the simplex and interior-point methods. We study non-linear programming (NLP) and its software through unconstrained optimisation methods and constrained optimisation methods. Instead of discussing dynamic programming (DP) directly, we discuss an important large-scale optimisation technique - decomposition. A number of definitions for heuristics are discussed and clarified in Chapter 3. Substitution is proposed as a principle which can be used to classify heuristics. Since many well-known heuristics can be categorised as problem substitution heuristics, model substitution heuristics, or algorithm substitution heuristics. The heuristic approach is then studied according to these classifications. The concept of computer simulation is discussed in Chapter 4. The method to categorise computer simulation according to the major objectives: problem-finding, problem-understanding, and problem-solving. In order to demonstrate these approaches, a case study using different approaches is proposed in Chapter 5. On the basis of the discussion on managerial decision making and an interview with the OR professionals, we carried out an extensive discussion on these three approaches through methodology consideration, modelling consideration, feasibility consideration, and applicability consideration. This research indicates that heuristics are more suitable for managerial decision making in many situations because optimisation approach still needs to be improved for managerial decision making. It is concluded that operations management should consider integrating these heuristic methods into a decision support system. Simple, understandable and usable heuristic approaches for solving managerial decision making problems are needed. Heuristics will be more attractive to managers in the future with the development of IT. Such approaches will undoubtedly be built around computer-based decision support systems. Computer simulation may then be regarded as the last resort. Despite this, it is surprising how often such an approach is needed. There are certain advantages in employing a simulation approach in management science and it may be the only way of tackling some managerial decision making problems. Diversified heuristics will be widely applied in managerial decision making.