Cloud Manufacturing Service Composition Optimization with Improved Genetic Algorithm

Abstract

Aiming at the problems in which there exists collocation between services and manufacturing tasks, multiobjective cloud manufacturing service composition optimization seldom considers the synergy degree of composite cloud services and the complexity of service composition, so a novel service composition optimization approach, called improved genetic algorithm based on entropy (IGABE), is put forward. First, the mathematical expressions of service collocation degree, composition synergy degree, composition entropy, and their related influence factors of the service composition are analyzed, and their definitions and calculation methods are given. Then, a multiobjective cloud manufacturing service composition optimization mathematical model is established. Moreover, crossover and mutation operators are improved by introducing normal cloud model theory and piecewise function, and improved roulette selection method is used to perform the selection operation. And the fitness function of the proposed IGABE is designed by combining Euclidean deviation with angular deviation. Finally, the manufacturing task of a wheeled cleaning robot is exemplified to verify the correctness of the proposed multiobjective optimization model for cloud manufacturing service composition and the effectiveness of the proposed algorithm, compared with Standard Genetic Algorithm (SGA), Hybrid Genetic Algorithm (HGA), and Cloud-entropy Enhanced Genetic Algorithm (CEGA). The studied results show that IGABE converges faster than SGA and HGA and can analyze and reflect the content difference expressed by the objective functions of service composition scheme and its approximation degree to the corresponding dimensions of the ideal point vector more comprehensively than CEGA. As such, the optimal service composition obtained by IGABE algorithm can better meet the complex needs of users.