A new algorithm using front prediction and NSGA-II for solving two and three-objective optimization problems

Abstract

In this paper, a new hybrid algorithm (FP–NSGA-II) is proposed by combining the fast and elitist non-dominated sorting genetic algorithm-II (NSGA-II) with a simple front prediction algorithm. Due to the significant computational time of evaluating objective functions in real life engineering problems, the aim of this hybrid approach is to better approximate the Pareto front of difficult constrained and unconstrained problems while keeping the computational cost similar to NSGA-II. FP–NSGA-II is similar to the original NSGA-II but generates better offsprings. This is achieved by using a prediction operator which utilizes the direction in the decision variable space between each solution in the first front and the nearest neighbour solution in the second front, in order to extrapolate future chromosomes. This enables the addition of solutions that are closer to the true Pareto front into the new generation. To assess the performance of the proposed approach, eight benchmark two-objective test problems and four three-objective test problems are used to compare FP–NSGA-II with NSGA-II. In addition, a three-objective heat exchanger network problem is examined to show the potential application of FP–NSGA-II in real-life problems. Results indicate that the FP–NSGA-II improves upon the performance of NSGA-II for a variety of benchmark test problems exhibiting different characteristics. In addition, a similar front prediction algorithm could also be easily integrated with other evolutionary algorithms to enhance its performance.