Parallel hybrid Price-genetic algorithm for global optimization

A hybrid algoritm to find the absolute extreme point of a multimodal scalar function of many variables is presented. The algorithm is well suited to be implemeted in parallel over a distributed computer environment. The algorithm is suitable when the objective function is expensive to compute, the computation can be affected by noise and/or partial derivatives cannot be calculated. The method used is a genetic modification of a previous algorithm based on the Price’s method. The genetic part of the algorithm is very effective to escape from local attractors of the algorithm and assures convergence in probability to the global optimum. The parallel implementation of the algorithm is based on a domain decomposition approach, fully exploiting available information over global behaviour of the function, reducing to a small amount the computer charge due to need for syncronization. The algorithm has been tested for optimizing several multimodal test function. Key-words:Global Optimization, Controlled Random Search, Genetic Algorithms, Parallel Computation.