The Optimization Issue Of Reliability And The Heuristic Algorithms In Wireless Sensor Network

Concerning about two optimization problems which are about the two-terminal reliability of computer communication networks and the number of the active sensor nodes in wireless sensor network, we mainly discuss the optimization problem of two-terminal reliability when some sensor nodes are work normally, and the optimization issue of the active sensor nodes' number under the restrictions of two-terminal reliability and the power of network. Two new algorithms for the optimization issues above are presented.1. It studies the reliability optimization about which the two-terminal reliability and topology of network will be affected because of the sensors' energy being exhausted in the wireless sensor network. This paper proposes the reliability optimization for edge-disjoint paths under m nodes' malfunction. By introducing the notion of s-t sub-graphs, a math model is set up. It presents a heuristic optimization algorithm, maximum edge-disjoint pathset reliability algorithm(MEDPR algorithm), which can figure out the most optimal s-t sub-network reliability as C≤C₀ that is m≤|V|-C(C is the number of active sensor nodes of the most reliable s-t sub-graphs gained from the MEDPR algorithm). The proposed algorithm is in polynomial time. Consequently simulation results illustrate the effectiveness of the algorithm. Furthermore, we also deal with the case that C ≥ C₀ in the optimization model, where C is the number of nodes in the shortest s-t path.2. This paper proposes the minimum number of active sensor nodes optimization issue subjects to the conditions of two-terminal network reliability and the power cost of a certain area. By introducing the notions of s-t sub-graphs and the α-area's power cost, a math optimuization model is set up. In this model, we present a heuristic optimization algorithm, minimum nodes algorithm subjects to reliability and power (MNRPalgorithm), which can figure out the minimum number of active sensor nodes as the sensor nodes satisfy the initial values in the constraint condition. We prove the proposed algorithm is in polynomial time also. Consequently simulation results illustrate the algorithm can solve this optimization problem effectively under the different initial values.