Improve micro genetic algorithm for multiobjective kursawe function and low pass filter circuit design optimization

Although Evolutionary Algorithms (EAs) have been widely implemented for solving Multiobjective Optimization Problems (MOPs), the convergence of EAs towards Pareto optimal front is still an issue of concern. In order to enhance the robustness of EAs, hybrid algorithms are commonly developed to identify better solutions for MOPs. The prime focus of this research is placed on the integration of new proposed elitism in conventional Micro Genetic Algorithm (MGA). The proposed elitism has been studied in this research to develop Improved Micro Genetic Algorithm (IMGA). In this research, Kursawe and ZDT test functions are chosen as the benchmark studies for the assessment on IMGA. The accuracy and effectiveness of IMGA are evaluated based a number of quality indicators such as generational distance and non-dominated optimal spacing. The proposed IMGA is compared with Non-dominated Sorting Genetic Algorithm II (NSGA-II), Strength Pareto Evolutionary Algorithm 2 (SPEA2), MGA and Fast Pareto Genetic Algorithm (FPGA). The assessment results show that IMGA can surpass the MGA in Kursawe test function by achieved 3.571E-4 for generational distance and 2.026E-2 for spacing. Meanwhile for ZDT benchmark, IMGA solved and suggested the optimal Pareto front for all the ZDT test functions. After having the benchmark evaluation, the proposed IMGA is applied to a practical case study on circuit design optimization. Two different circuit designs of active low pass filter that comprise of different number of input parameters are studied.