System Optimization Using a Parallel Stochastic Approach

This paper describes an original stochastic algorithm based on a parallel approach. The algorithm is suitable especially for a real technical system optimization. A few independent pseudorandom generators are used. They generate independent variable vectors along all of the optimized system axes. Local optimal values are used to define a final pseudorandom generator with a narrower interval around the global optimum. Theoretical foundations are introduced and a few practical experiments are presented. The described method is also suitable for the quality classification of the pseudorandom generators using the selected RGB color scheme. Main advantages of this approach are discussed. The algorithm was developed in the MATLAB environment.