PARALLEL HYBRID METHODS USED IN OPTIMIZATION PROBLEMS SOLVING

Abstract

This paper presents different models of hybrid algorithms that can be run on parallel architectures being used in optimization problems solving. In these models we used several techniques: genetic algorithms, ant colony and tabu search. Optimization problems can achieve a high degree of complexity, which is the main reason for the necessity of using of these methods in such incursions. With their cooperation, we tried to obtain satisfactory results in much better running time than the sequential versions. These models have been run using various parallel configurations  on a cluster cores, which  belong to „Stefan cel Mare” University. The results obtained for these models were compared with each other and with the results obtained for models described in other personal papers. The paper highlights the advantages of the parallel hybrid cooperation in solving of complex optimization problems.  This paper is structured in four chapters: Introduction, Cooperative heterogeneous model, Cooperative hybrid models and Conclusions.