Abstract:
Many real-world chemical industrial problems involve two types of problem difficulty: i) multiple conflicting objectives (Multi-objective optimization) and ii) a highly complex search space. Thus, efficient optimization strategies being capable of solving problems with both types of problem difficulty are important. Evolutionary algorithm (EA) can efficiently solve multi-objective optimization problem with highly complex search space. Unfortunately, EA requires many parameters. Consequently, the selection of the parameters value is important, and has an effect on accuracy and convergence of the solution obtained. This research focuses on analyzing the effect of genetic parameters (mutation and crossover probability) in evolutionary algorithm for multi-objective optimization problems. In this work, genetic parameters for MOGA, NSGA, NPGA, NPGA, NSGA-II and SPEA are investigated. Furthermore the generic guideline to select suitable parameters values for the multi-objective evolutionary algorithms is developed and applied in a case study. The case study involves the problem of synthesis of phenol recovery process. The objective of this problem is to find the suitable operating points providing a minimum total cost and minimum environmental impact. The generic guideline for selection appropriate genetic parameters is successfully applied to the case study.
