A Novel Physics Inspired Multi-objective Optimization Algorithm: Multiple Objective Gravitational Optimization
Rajdeep Chatterjee, Madhabananda Das · 2015
This paper proposes a new multi-objective optimization algorithm inspired by the Newtonian Law of Gravity. The new algorithm is a multiple objective extension of the single objective optimization Gravitational Search Algorithm. Generally, we know various multi-objective algorithms with leader selection schemes. The Leader guides the population towards Pareto fronts. As a result, the algorithm becomes complex in nature. To reduce that part of the cost, we introduce a new algorithm which does not use the leader for guidance but relies on its own population to obtain the non-dominated set of solutions. The algorithm has been tested on several benchmark functions and has produced a good number of solutions. Our purpose of research is to showcase that the memory-less property of the single objective Gravitational Search Algorithm can be used to solve multiple objective problems.