Collaborative Co-evolution Algorithm And Its Application

Co-evolutionary Algorithm (CA) is one algorithm which has emerged during recent years. Although its development time was not long, the algorithm has been successfully used in some optimization question, including many NP questions. It comes from the traditional evolution algorithm, also shows its own characteristics: in CA, individuals existing mutually cooperation relations between different species groups, instead of the only competition relations in the traditional evolution algorithms. At present, a lot of work about CA has been done by the scholars both abroad and at home.In this paper, fistly, we analyzed the research works that scholars are going along with in the aspect of CA. Secondly, the core and basic route of this algorithm were introduced; In order to overcome the premature convergence problem which appears in the evolution process, we proposed a new algorithm called Cooperative Coevolutionary Genetic Algorithm Based on Chaos (CCGA-G); It used the chaos principle to break the monopoly phenomenon of partial outstanding individual, and caused the individual in the evolution population widely distributed. Finally, we applied the cooperative coevolutionary genetic algorithm to practical problem, e.g. production scheduling problem and the problem of data clustering. In production scheduling applications, the job shop scheduling problem has been analyzed in detail, and an encoding method of co-evolution that fits job shop scheduling has been proposed. In data clustering applications, the paper linked up data clustering with graph partition to keep the question as brief as possible and easy to solve.