The Measure of Pareto Optima Applications to Multi-objective Metaheuristics
This article describes a set function that maps a set of Pareto optimal points to a scalar. A theorem is presented that shows that the maximization of this scalar value constitutes the necessary and sufficient condition for the function’s arguments to be maximally diverse Pareto optimal solutions of a discrete, multi-objective, optimization problem. This scalar quantity, a hypervolume based on a Lebesgue measure, is therefore the best metric to assess the quality of multiobjective optimization algorithms. Moreover, it can be used as the objective function in simulated annealing (SA) to induce convergence in probability to the Pareto optima. An efficient, polynomial-time algorithm for calculating this scalar and an analysis of its complexity is also presented.
KeywordsObjective Function Simulated Annealing Multiobjective Optimization Pareto Optimal Solution Pareto Optimum
Unable to display preview. Download preview PDF.
- 1.Fleischer, M.A.: The measure of pareto optima: Applications to multiobjective metaheuristics. Technical Report 2002-32, Institute for Systems Research, University of Maryland, College Park, MD. (2002) This is an unabridged version of the instant article.Google Scholar
- 2.Zitzler, E., Thiele, L.: Multiobjective optimization using evolutionary algorithms—a comparative case study. In A.E. Eiben, T. Bäck, M.S.H.S., ed.: Proceedings of the Fifth International Conference on Parallel Problem Solving from Nature—PPSN V, Berlin, Germany (1998)Google Scholar
- 3.M. Laumanns, G.R., Schwefel, H.: Approximating the pareto set: Diversity issues and performance assessment. (1999)Google Scholar
- 4.Zitzler, E.: Evolutionary Algorithms for Multiobjective Optimization: Methods and Applications. Ph.d. dissertation, Swiss Federal Institute fo Technology (ETH), Zurich, Switzerland (1999) The relevant part pertaining to our Lebesgue measure is discussed in Ch. 3.Google Scholar
- 6.Fonseca, C., Fleming, P.: On the performance assessment and comparison of stochastic multiobjective optimizers. In G. Goos, J.H., van Leeuwen, J., eds.: Proceedings of the Fourth International Conference on Parallel Problem Solving from Nature—PPSN IV, Berlin, Germany (1998) 584–593Google Scholar
- 8.M. Laumanns, L. Thiele, K.D., Zitzler, E.: Archiving with guaranteed convergence and diversity in multi-objective optimization. In: Proceedings of the Genetic and Evolutionary Computation Conference (GECCO). (2002) 439–447Google Scholar
- 9.Coello, C.: Bibliograph on evolutionary multiobjective optimization. http://www.lania.mx/~coello/EMOO/EMOObib.html (2001)
- 15.Wu, J.: Quality Assisted Multiobjective and Multi-disciplinary Genetic Algorithms. Department of mechanical engineering, University of Maryland, College Park, College Park, Maryland (2001) S. Azarm, Ph.D. advisor.Google Scholar
- 16.Knowles, J.: Local Search and Hybrid Evolutionary Algorithms for Pareto Optimization. PhD thesis, The University of Reading, Reading, UK (2002) Department of Computer Science.Google Scholar
- 20.Fleischer, M.A.: 28: Generalized Cybernetic Optimization: Solving Continuous Variable Problems. In: Metaheuristics: Advances and Trends in Local Search Paradigms for Optimization. Kluwer Academic Publishers (1999) 403–418Google Scholar