بهینه سازی پارادایم الگوریتم یادگیری اجتماعی (SLO) و کاربرد آن در ترکیب سرویس های آگاه QoS ابری

Inspired by the evolution process of human intelligence and the social learning theory, a new swarm intelligence algorithm paradigm named the social learning optimization (SLO) algorithm is proposed. SLO consists of three co-evolution spaces: the bottom is the micro-space, where genetic evolution occurs; the middle layer is the learning space, where individuals enhance their intelligence through imitation learning and observational learning; knowledge is extracted from the middle layer and delivered to the top layer, which is called the belief space, where culture is established through knowledge accumulation and used to guide individuals’ genetic evolution in the micro-space regularly. SLO is an optimization algorithm model for optimization problems, and a concrete algorithm could be generated by embodying SLO's three evolution spaces. Moreover, to demonstrate how to employ SLO and verify its superiority, this paper proposes the specific SLO (S-SLO) to solve the problem of QoS-aware cloud service composition. S-SLO is constructed by integrating the improved differential evolutionary (DE) algorithm and improved social cognitive optimization (SCO) into the micro-space and the learning space, respectively. Finally, experimental results and performance comparison show that the S-SLO is both effective and efficient. This work is expected to explore a novel swarm intelligence optimization model with better search capabilities and convergence rates, as well as to extend the theory of the swarm intelligence optimization algorithm.