Comprehensive Analysis of Stochastic Algorithms: An Alternative in Optimization

Global optimization is an important branch of computational mathematics that finds the applications inevery walk of life. The real-world science and engineering optimization applications are becoming moreand more complex in nature and are generally multimodal. The conventional optimization methods fail tooptimize such complex multimodal problems. Hence there is always an increasing demand for efficient androbust optimization strategies. In recent years, algorithms based on random process have become popularand are an alternative to conventional methods in optimization. The Stochastic Algorithms are based onrandom process and most of them are the extraction of natural phenomenon for problem solving. Theobjective of this paper is to investigate and analyze stochastic algorithms on complex multimodaloptimization problems. The comprehensive analysis of different stochastic algorithms is carried out on aset of standard benchmark problems with 10, 30 and 50 dimensions. The algorithmic suitability, robustnessand convergence rate of each will be investigated. Finally the dependency of Stochastic Algorithms onproblem dimensions are also discussed..