Research on QoS service composition based on coevolutionary genetic algorithm

Traditional genetic algorithms overemphasize the struggle for survival and neglect all other aspects of biology. In addition, binary encoding is widely used in individual coding. Since the individual chromosomes produced are longer in length, it is difficult to ensure the efficiency of the algorithm. In this study, a coevolutionary genetic algorithm is proposed for web service composition based on quality of service (QoS), which fully considers the individual relationships among populations. The real coding method is adopted to solve the service selection problem based on QoS, so that the negative effect of the long length of chromosomes in the algorithm is avoided. Moreover, in view of the difficulty of determining the weight of each QoS attribute in web services, we propose to use the entropy method to determine the weights of each one. Compared with the traditional genetic algorithm, the experimental results show that the proposed algorithm converges faster in the service composition, and the fitness of the optimal solution is higher.