Non-Linear Precomputation For Optimal Data sources and Paths Searching

While data and path redundance avoids, to some extent, data damages and mission failures resulting from node failures in networks, users must face up to the great challenge of Quality of Service(QoS), i.e., how to select optimal data sources and paths among different data sources. We address ourselves to the problem of Multiple Data Sources Selection(MDSS) for data sharing and propose a precomputation algorithm, namely PAMDSS. PAMDSS decomposes MDSS into two sub-problems, and introduces the concept of Pareto optimization which reduces the search space greatly. By means of nonlinear path length based precomputation, PAMDSS achieves good QoS effects. Extensive simulations show the efficiency of our algorithm.