A View of Costed Pi-calculus for Particle Swarm Optimization with QoS-aware Service Selection

摘要

This paper presents a novel computation model for Particle Swarm Optimization (PSO) based on process algebras, which consists of some additional primitives, such as beam channel, logic parallels, logic choices, cost, cost choices and so forth. In this calculus, everything is a cost expression including the cost of a expression, the choice of expressions, or even the search procedure of PSO itself. As an instance of the applications using discrete PSO, QoS-aware web service selection can be modeled naturally in this calculus and the algebraic and physical representation of the solution is imported to guideline the design of discrete scheme of PSO. In the experiments in simulation platform, the results show that the proposed approach can effectively formalize the QoS-aware web service composition in both algebraic and physical perspectives, and also generate the well-scrutinized service in accordance with the high QoS cost requirements in the acceptable time.