NODE DEPLOYMENT FOR MAXIMIZATION OF COVERAGE AREA IN WIRELESS SENSOR NETWORK

Abstract

A Wireless Sensor Network is a collection of autonomous sensors that work in coordination with each other to achieve a predetermined objective. Today wireless sensor networks are being used in a gamut of application belonging to both civilian and military domains. One of the open problems with respect to wireless sensor networks is the problem of finding an optimal deployment sequence that covers the entire mission area so that the best quality of service can be provided with low cost. In this dissertation, two techniques for dynamic node deployment based on the Artificial Bee Colony (ABC) algorithm in wireless sensor networks are proposed. The first proposed algorithm incorporates two additional phases namely the crossover phase and the opposition based learning phase apart from those of the ABC algorithm. The crossover phase is used for replacing a certain amount of worse solutions with new solutions that were generated using multi-point crossover operation. The opposition based learning phase generates an opposite solution for each of the F food sources, hence doubling the number of available food sources to 2F after which we selected the best F food sources out of them. The opposition based learning phase was carried out only 30% of the time. The second proposed algorithm uses a hybrid local search for striking a balance between the exploration and exploitation behavior of the algorithm. The hybrid local search constituted of three search strategies namely the perturbation based local search strategy of ABC, the global best-directed search and a variable step size local search. The algorithm uses all the three strategies for finding a new solution around an old solution and makes a greedy selection amongst the three. This algorithm too like the previous one uses a crossover phase to increase the probing performance of the algorithm. Both of the proposed algorithms were simulated, and results were compared with the ABC algorithm, it was observed that both the algorithms outperformed the ABC algorithm in finding a deployment sequence for maximizing area coverage when executed under similar conditions.