Single and multi-objective optimization - a comparative analysis

Abstract

We compare the order relations used in single and in multi-objective optimization. The single objective optimization is based on a unique complete ordering. In multi-objective case there are infinite inequivalent partial orderings. The implications of these facts are analysed with examples. Some implications on Pareto techniques are also pointed out. A technique will be presented on the use of optimal frontier which synthesises it into a "compiled" guide to control the system successfully. Some applications as examples are presented like economic planning strategies etc.