Abstract:In optimization, multiple objectives and constraints cannot be handled independently of the underlying optimizer. Requirements such as continuity and differentiability of the cost surface add yet another conflicting element to the decision process. While "better" solutions should be rated higher than "worse" ones, the resulting cost landscape must also comply with such requirements. Evolutionary algorithms (EAs), which have found application in many areas not amenable to optimization by other methods, possess many characteristics desirable in a multiobjective optimizer, most notably the concerted handling of multiple candidate solutions. However, EAs are essentially unconstrained search techniques which require the assignment of a scalar measure of quality, or fitness, to such candidate solutions. After reviewing current revolutionary approaches to multiobjective and constrained optimization, the paper proposes that fitness assignment be interpreted as, or at least related to, a multicriterion decision process. A suitable decision making framework based on goals and priorities is subsequently formulated in terms of a relational operator, characterized, and shown to encompass a number of simpler decision strategies. Finally, the ranking of an arbitrary number of candidates is considered. The effect of preference changes on the cost surface seen by an EA is illustrated graphically for a simple problem. The paper concludes with the formulation of a multiobjective genetic algorithm based on the proposed decision strategy. Niche formation techniques are used to promote diversity among preferable candidates, and progressive articulation of preferences is shown to be possible as long as the genetic algorithm can recover from abrupt changes in the cost landscape.

A new method for decision making in multi-objective optimization problems

A Hybridized Vector Optimal Algorithm for Multiobjective Designs of Inverse Problems

Multiobjective optimization and multiple constraint handling with evolutionary algorithms. I. A unified formulation

A novel multiobjective evolutionary algorithm based on decomposition and multi-reference points strategy

A Reference Point-Based Evolutionary Algorithm Solves Multi and Many-Objective Optimization Problems: Method and Validation

Evolutionary Many-Objective Optimization of Hybrid Electric Vehicle Control: from General Optimization to Preference Articulation

Interactive Evolutionary Multiobjective Optimization Via Learning to Rank

Applying Social Choice Theory to Solve Engineering Multi-objective Optimization Problems

A Tradeoff-Based Interactive Multi-Objective Optimization Method Driven by Evolutionary Algorithms.

Multiobjective evolutionary algorithms: a comparative case study and the strength Pareto approach

A Hybrid Framework for Evolutionary Multi-Objective Optimization

An Evolutionary Algorithm with Advanced Goal and Priority Specification for Multi-objective Optimization

Multiobjective Optimization of Non-Smooth PDE-Constrained Problems

MULTIOBJECTIVE DESIGN OPTIMIZATION BY AN EVOLUTIONARY ALGORITHM

An efficient multi-objective optimization algorithm based on swarm intelligence for engineering design

Multi-modal Multi-objective Optimization: Problem Analysis and Case Studies

A surrogate-assisted a priori multiobjective evolutionary algorithm for constrained multiobjective optimization problems

An Adaptive Switching-Based Evolutionary Algorithm for Many-Objective Optimization

Multi-objective optimization algorithms applied to the class integration and test order problem

Knee based multimodal multi-objective evolutionary algorithm for decision making

An interactive fuzzy multi-objective optimization method for engineering design