Performance Of Differential Evolution And Particle Swarm Methods On Some Relatively Harder Multi-modal Benchmark Functions

Global Optimization (GO) is concerned with finding the optimum point(s) of a non-convex (multi-modal) function in an m-dimensional space. Although some work was done in the pre-1970's to develop appropriate methods to find the optimal point(s) of a multi-modal function, the 1970's evidenced a great fillip in simulation-based optimization research due to the invention of the 'Genetic algorithm' by John Holland (1975). All stochastic search methods of global optimization partake of the probabilistic nature inherent to them. As a result, one cannot obtain certainty in their results, unless they are permitted to go in for indefinitely large search attempts.