Abstract
Network reliability is very important for the decision support information. Monte Carlo Simulation (MCS) is one of the optimal algorithms to estimate the network reliability for different kinds of network configuration. This thesis has compared and analyzed three Monte Carlo simulation (MCS) methods for estimating the two-terminal network reliability of a binary-state network: (1) MCS1 simulates the network reliability in terms of known MPs, (2) MCS2 estimates the network reliability in terms of known MCs; and (3) MCS3 estimates the network reliability directly without knowing any information of MPs or MCs. Our simulation results show that the direct estimation without knowing any information of MPs or MCs can speedup about 195 times when compared with other traditional approaches which require MPs or MCs information. In addition, we also combine particle swarm optimization (PSO) and MCS to solve cost minimization problem under reliability constraints. Compared with previous works to solve this problem, the result of PSO combine with MCS can get the better solution.