Multiobjective models for business decision making: Applications to production planning and investment analysis

A characteristic of most formal techniques that have been used for business decision making is the selection of the best course of action with respect to certain judgmental criterion. However, the nature of many decision problems has changed considerably in recent years and, in most practical situations, the management often encounters an array of conflicting goals in that the achievement of some is usually at the expense of others. In light of this trend in business decision making, this research examines the use of multiobjective approaches to solve a number of problems encountered by managers from companies of dissimilar nature. Essentially, the dissertation is composed of three studies, each representing an application of multiobjective decision models for a specific environment, namely, a manufacturing job shop where the characteristics and requirements of customer orders are not predictable, a two-stage production flow line found in cellular manufacturing system or Just-In-Time environment, and a financial service (or asset management) company whose primary goal is maximizing wealth while taking into consideration various types of risk.; The first essay examines the influence of order arrival pattern in a job shop using computer simulation. The issue of batching of customer orders is also investigated. A MANOVA statistical experiment is used to analyze the simulation outputs a multiobjective standpoint. Results suggest that order arrival pattern affects the choice of dispatching rule and other environmental factors. In the second essay, we develop three different sets of heuristic procedures to solve a bicriteria scheduling problem for a two-stage flowshop. The objectives in consideration are minimization of total flowtime and number of tardy jobs. Results of the multiobjective problem are a Pareto set of schedules which can be used for further analysis. In addition, the study provides enhancements to a variety of lower bounds with respect to flowtime and number of tardy jobs in a flowshop environment. The third essay in this dissertation investigates a multiobjective approach to combine analysts' forecasts. The approach makes use of a goal programing model to determine the optimal way to construct a composite investment portfolio based on the distributional properties of the historical return of analysts' forecasts. Empirical experiments show that this multiobjective approach performs better than other forecast-combining methods subject to various investor preferences.