Research On Theory And Application Of Improved Particle Swarm Optimization Based On Cloud Model

Many scientific, engineering and economic problems need the optimization of a set of parameters with the aim of minimizing or maximizing the objective functions. Particle Swarm Optimization (PSO) is a kind of group intelligent heuristic global evolutional algorithm that was proposed in the recent decade. With the advantages such as its principle is simple, it has fast convergence rate and it is easy to be implemented, it is applied to solve non-linear and non-differentiable complex optimization problem, also it can be used to solve combinatorial optimization problem too. PSO has become an important branch in artificial intelligence research areas in recent years. However, PSO has some inherent defects such as following:premature convergence, easily turning into local optimization and poor search result. In response to these defects, taking steps to improve PSO algorithm and expand application fields will have important theoretical value and practical significance.On the base of systematic analysis on basic theory and general improved principles of PSO algorithm, and integrating with the out standing characteristics of the cloud model on the process of transforming a qualitative concept to a set of quantitative computation, the authors propose a novel rapid evolutionary algorithm, Cloud Particle Swarm Optimization algorithm. It includes two forms CCPSO and CHPSO. With the cloud model, inheritance and mutation of particle can be modeled naturally and uniformly, which make it easy and nature to control the scale of searching space. This enables the algorithm to be able to find accurate numerical solutions within a short time. Operator though normal cloud particle operator to achieve the evolution of the mutation operation, and can overcome the premature problem validly. The simulation results show that the algorithm has better probability of finding global optimum, especially for multimodal function.Hard ware/Software partitioning is one of the key problems in embedded system co-design. It provides the best partitioning scheme for the Embedded System in the multi-constraint condition. Based on the characteristics of the cloud model on the process of transforming a qualitative concept to a set of quantitative numerical values, this paper describes an approach based on immune particle swarm optimization algorithms. Clone copy operator, clone hypermutation operator and clone selection operator are redefined. Experimental results show that the algorithm improve the precision of the optimal solution and get reasonable partitioning results.