Comparison between Multiobjective Population-based Optimization Algorithms in Mechanical Problem
The objective of this paper was to perform a comparative study among multi-objective optimization methods on practical problem by using modeFRONTIER optimization software, to determine the efficiency of each method. In order to measure the effectiveness and competence of each method, the lifting arm problem was chosen from the literature. Two numerical performance metrics and one visual criterion were chosen for qualitative and quantitative comparisons:(1) the variance of solution distribution in the Pareto optimal regions, (2) the ratio between the number of resulting Pareto front members to total numbers fitness function calculations which is denoted by hit rate, and lastly (3) graphical representation of the Pareto fronts for discussion. These metrics were chosen to represent the quality, as well as speed of the algorithms by ensuring well extends solutions.
The definition of the variance as the sum of the square difference between the distance of each Pareto solutions and the average distance between Pareto solutions, over the total number of Pareto solutions. Comparisons among the results obtained using different algorithms have been performed to verify their performance. The experiments carried out indicate that FMOGA-II obtains remarkable results regarding all metrics used.