Study On Free Search Optimization And Its Applications To Sensor Network

Swarm intelligence is one of the artificial intelligence approaches inspired by the intelligent phenomenon of the social animals just like ant colony, school of fish, bird flock in nature. Due to the characteristics of distributed control, overall information dissemination and self-organizing, swarm intelligence has become an important computation intelligence branch, and is widely used in the fields of computer science, intelligent control, communications networking and robotics et al.. Its excellent performance and potential in the application make it a research hotspot in the field of intelligent computing. This paper makes an investigation on a new swarm intelligence optimization--Free Search algorithm (FS), and further extends Free Search algorithm to the application of node localization estimation and coverage control in wireless sensor network (WSN). The main research work in the paper is organized as follows.1) The extensive studies of swarm intelligence in the literature are reviewed including the origin, main characteristics and the advantages of swarm intelligence. And then Free Search algorithm is further investigated. Not only the working.mechanism, model and implementation of the algorithm are explored, but also the effectiveness of the algorithm is verified by the simulation results. In addition, Free Search algorithm is compared with particle swarm optimization (PSO) from their biological backgrounds, the search mechanism, the simulation results and other aspects. And then the research trends of Free Search algorithm are pointed.2) To overcome the deficiencies of Free Search algorithm (especially the problem of the sensitive individual neighborhood search radius) and to improve the environment adaptability and robustness of the algorithm, reach the equilibrium point of global search and local search, the adaptive strategy of the algorithm is investigated and two improved algorithms are proposed:i) the improved Free Search algorithm (IFS); ii) the adaptive search algorithm (AFS). Simulation results indicate that the two improved algorithm can not only overcome the shortage of the original algorithm, effectively avoid local sub-optimal, but also have a higher convergence speed, convergence precision and optimization success rate than the former one.3) After further studying the hybrid strategy, a hybrid algorithm of genetic algorithm (GA) and Free Search, named genetic free search algorithm (GAFS) is proposed. Two new operators named coarse-grained crossover and fine-grained crossover are introduced in GAFS. The coarse-grained crossover operator among individuals makes the algorithm obtain strong global exploring ability, and the fine-grained crossover operator within the individual search radius improves the local search ability. As a result, the algorithm can get the dynamic equilibrium between exploration and exploitation.4) Two operators (coarse-grained crossover and cloning mutation) are defined to further explore the role of hybrid strategy. The idea of GA is integrated into the particle swarm optimization, and then a new genetic particle swarm optimization algorithm (GAPSO) is proposed. Coarse-grained crossover is used to overcome the phenomenon of particles'aggregation in the evolution process, enable the algorithm to escape from local sub-optimization effectively and to improve search success ratio. Cloning mutation is used to solve the problems that the PSO algorithm slows down convergence in later stage of evolution and barely meets the precision requirements. Thus the search capability and efficiency of the algorithm are improved.5) Based on the studies on elite-based multi-objective optimization model, the population diversity strategies, non-dominated sorting and the fitness assignment about multi-objective evolution, FS is applied to multi-objective optimization. As a result, Multi-Objective Free Search algorithm (MOFS) comes out. During the design process of MOFS, the main contributions are as follows:â…°)a new method to define fitness function is proposed;â…±) a new search mechanism is constructed by redefining the pheromone and sensitivity;â…²) the adaptive strategy is proposed;â…³) the external archive is established with the grid-based diversity preservation strategy. In addition, simulation results verify the correctness and efficiency of the MOFS algorithm. 6) FS and MOFS are applied to node localization estimation in wireless sensor network(WSN) and coverage control of WSN respectively. Node localization is one of the most important issues in WSN after the deployment, because without location conditions the information of node would be meaningless. Thus, the FS-based intelligence estimate localization algorithms is proposed by introducing FS to solving node localization problem in wireless sensor network and to obtaining the best estimated coordinate of unknown node.Coverage control is a basic problem in WSN applications. The feasible coverage control strategy can improve the quality of service (QoS) and extend the network lifetime of WSN. This paper proposes WSN coverage control multi-objective optimization model to optimize network coverage and network lifetime of WSN, which turns WSN coverage control problem into the multi-objective optimization problem. During the network operation, WSN is real-time optimized by the multi-objective free search algorithm to obtain the optimal set of active nodes, so that the network always works in the optimal working condition.