An Improved Particle Swarm Optimization Algorithm For Solving Leakage Detection Problem

Further development of modern science has promoted researches on source localization. Using artificial intelligence techniques to solve source localization problem has urgent needs in many areas such as industrial areas, production areas and living areas.The problem of source localization consists in finding, by one or several mobile or fixed sensors arrays, possibly cooperating with each other, the point or the spatial region from which a quantity of interest is being emitted. Because source localization is closely related to production and living areas, it is valuable enough to be researched. In this paper, a branch of the source localization field-gas leaking detection field is further studied. Different from previous methods such as using pressure principle or non-interactive sensors to detect, or using complex source localization algorithms, this paper views from a new perspective, which is making use of Swarm Intelligence optimization algorithms, and combining with Multi-Agent Technology to solve the gas leaking detection problem.Swarm Intelligence optimization algorithms study is a hot issue of Artificial Intelligence. The study appeared in 1990s, caused great concern of many scholars. Swarm Intelligence optimization algorithms originated from common cluster behaviors in nature. The famous Swarm Intelligence algorithms include Ant colony optimization algorithm and Particle Swarm Optimization. The features of Swarm Intelligence optimization algorithms include distributed control, high scalability, simple rules, easy realization and self-organization, etc.Based on ideas above, this paper explores the possibility of solving gas leaking detection problem using PSO algorithm. Facing the communication radius limit of Agents, aiming at the defects and deficiencies of local PSO algorithm, this paper tries to improve the algorithm on the basis of strategy of Wireless Sensor Networks Coverage to make the algorithm more realistic.In this paper, the main work and innovations are as follows:(1)This paper attempts to apply PSO algorithm to solve gas leaking detection problem. As a branch of the source localization field, gas leaking problem is solved mainly by pressure principle or sensor without interaction for now. With the development of intelligent optimization algorithms and Multi-Agent technology, this paper generates the idea of using Agents which are simple and inexpensive to solve gas leaking detection problem.(2)Aiming at the limits of communication radius, this paper makes use of local PSO algorithm, and conducts simulation experiment. Because of communication radius, it is hard to get global optimum values for Agents. Although PSO algorithm can converges to the leaking points mostly, this paper learns to use local PSO algorithm to make the algorithm more realistic.(3)This paper improves the algorithm by combining with strategy of Wireless Sensor Networks Coverage, and puts forward a double-local PSO algorithm. Although local PSO algorithm can converges to the leaking points mostly in simulation experiment, it has exposed some questions in convergence rate and divergence probability. Aiming at this problem, according to strategy of Wireless Sensor Networks Coverage, this paper improves its initialization and then puts forward a double-local PSO algorithm. This paper conducts simulation experiments and verifies the improved results.