Chun-mei Chen, Bin Wu, Xian-guo Tuo, Hong Jiang, and Juan Zhang
multi-hop ad hoc networks, state pruning, metropolis criterion, relay
In this paper, we consider the multi-hop cooperative ad hoc networks with dynamic network status, limited node energy, and fading channel environment. It is difficult to solve the multi-objective cost function with large state space. We aim to find the optimal policies to reduce these states for arbitrary topologies with links affected by path loss. Therefore, this paper presents the algorithm of state pruning program based on inverse probability of metropolis (SPP_IPM) which defines the inverse probability of metropolis as the refuse acceptance criteria and realizes the pruning for state space. It achieves the pruning upper bound by scaling method and reduces the number of states according to a certain probability bound. Thus, we reduce the computation complexity through this pruning technique. The simulation results demonstrate SPP_IPM is better than other typical algorithms such as RTDP (real-time dynamic programming) and FRTDP (focused real-time dynamic programming) with respect to the state space, the success probability, and the transmission delay. Furthermore, SPP_IPM has a good control effect on the growth of the network state space when the network scene changes.
Important Links:
Go Back
