Constrained Multi-objective Optimization Using Steady State Genetic Algorithms

Abstract

Abstract. In this paper we propose two novel approaches for solving constrained multi-objective optimization problems using steady state GAs. These methods are intended for solving real-world application problems that have many constraints and very small feasible regions. One method called Objective Exchange Genetic Algorithm for Design Optimization (OEGADO) runs several GAs concurrently with each GA optimizing one objective and exchanging information about its objective with the others. The other method called Objective Switching Genetic Algorithm for Design Optimization (OSGADO) runs each objective sequentially with a common population for all objectives. Empirical results in benchmark and engineering design domains are presented. A comparison between our methods and Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) shows that our methods performed better than NSGA-II for difficult problems and found Pareto-optimal solutions in fewer objective evaluations. The results suggest that our methods are better applicable for solving real-world application problems wherein the objective computation time is large.

Asynchronous Evolutionary Multi-Objective Algorithms with heterogeneous evaluation costs

“…This confirms the results from [8], [9] demonstrating that steady-state algorithms performed often better than generational ones in several contexts, and in particular in a multiobjective setting. The situation is typically visible on Figure 4: the right plot shows the delay for NSGA-II epsilon , whereas the left plot illustrates the well-known benefit of asynchronous steady-state selection, in terms of elapsed time.…”

“…In a general real-world framework, [8] argues that steady-state performs very often better than generational, and even more so in a multi-objective optimization context. In [9], another comparison between steady-state and generational NSGA-II is proposed on a real case study.…”

Multi-objective evolutionary simulation-optimisation of a real-world manufacturing problem

“…Real-world optimisation problems often involve an immense number of possible solutions, and an EA needs a large number of simulation evaluations before an acceptable solution can be found [3]- [4]. Even with improvements in computer processing speed, one single simulation evaluation may take a couple of minutes to hours or days of computing time [4]- [5]. This poses a serious hindrance to the practical application of EAs in real-world scenarios.…”

A parallel surrogate-assisted multi-objective evolutionary algorithm for computationally expensive optimization problems

“…This algorithm is based on a steadystate design to support a high degree of parallelism. Most multi-objective EAs use a generational approach [5], which is not optimal with respect to parallel efficiency. In generational algorithms, results for a complete generation must be awaited in order for the search to proceed.…”

“…Real-world optimization problems often involve an immense number of possible solutions, and an EA needs a large number of simulation evaluations before an acceptable solution can be found [2]- [3]. Even with improvements in computer processing speed, one single simulation evaluation may take a couple of minutes to hours or days of computing time [4]- [5]. This poses a serious hindrance to the practical application of EAs in real-world scenarios, and to tackle this problem various approaches have been suggested.…”