A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering of the Asian Institute of Technology, Bangkok, Thailand

Thesis (M.Eng.) - Asian Institute of Technology, 1975

This thesis considers the two common cases of multicriterion optimization problems: the goal-attainment type and the type. a The classical approach to solving the first type is to set targets priori and use goal programming to achieve them. Its failure to provide a framework for setting targets necessitates model simulation through target variation resulting in tedious calculations. The SCOC (Sequential Conversion of Objective into Constraints) approach is an iterative procedure for rationally organizing the investigation of these acceptable and permissible solutions as a function of the partial choices made by the decision maker at the interactive stages of the process. Two sensitivity analyses - Consequence Realization and Behavior Tracing provided the grounds in undertaking the investigation. The second type of problem arises when there is absolute obscurity on the decision maker's preferences of the objectives. Existing inter active approaches fail in giving the decision maker enough information on the set of permissible solutions, thus requiring numerous interactions and tedious computations before an acceptable solution is arrived. The "indifference map" approach involves a systematic method of simulating all permissible "best compromise" solutions and presenting them in a manner which would facilitate the search for an acceptable solution. It relies on the concepts on sequential optimization and efficient points. This approach is extended to handle multicriterion quadratic programming problems.