A hybrid algorithm based on multi-agent and simulated annealing

Abstract

Recently, a number of approaches are introduced to improve the evolutionary computation and searching efficiency. Many of those which combine evolutionary algorithms and searching algorithms are witnessed to be efficient and important. In current research status of hybrid evolutionary algorithms, the optimal algorithms introduced in EA normally acts as local searching solution. On the other hand, when works with global searching of generic algorithms, the combination is considered a better approach for finding global optima. For example, an embedded steepest descent in generic algorithms can speed up the convergence of finding the global optima. It is achieved through adopting fast local searching and paralell searching of generic algorithms. The reported optimal algorithms that have been adopted in generic algorithms are mainly steepest descent algorithm, simplex algorithm and simulated annealing algorithm. These algorithms are recognized helpful to increase the performance of global searching in generic algorithms. This paper describes a Hybrid Evolutionary Algorithm (HEA) abstracting the advantages of both evolutionary algorithms and simulated annealing algorithms and then shows the simulation results of its performance. © 2012 Springer-Verlag GmbH.