Two-stage hybrid genetic algorithm for robot cloud service selection

Abstract Robot cloud service platform is a combination of computing and robotics, providing intelligent services for many robots. However, to select that satisfys the robot’s requirements from massive with different QoS indicator in an NP hard problem. In this paper, based on cost model between platform, we propose two-stage selection strategy, namely, candidate stage according specific robots final optimization stage. Additionally, respect optimizing model, Dynamic Vector Hybrid Genetic Algorithm (DVHGA) integrated local global search process as well three-phase parameter updating policy. Specifically, inspired by momentum deep learning, dynamic vector DVHGA modify weights ensure reasonable allocation resources. Moreover, suggest linear evaluation method concerning time at same time, which could be expected used real application environment. Finally, empirical results demonstrate proposed outperforms other benchmark algorithms, i.e., DABC, ESWOA, GA, PGA GA-PSO, convergence rate, total score.