Lifeline system networks include cities or regions’ water supply systems

gas transmission systems

electrical power systems

transportation systems and so on.The aim of networks’ seismic reliability analysis is used to guide lifeline network systems’ layout and design.Networks’ optimization problem usually involves a lot of factors and represents high non-linear character

so traditional optimization methods can’t solve the problem.With the development of computer techniques

modern heuristic optimization methods

such as tabu search method

simulated annealing algorithm

genetic algorithm and neural networks

are advanced and used widely to solve NP-hard problems.In this article

undirected edge-weight networks are analyzed.With network’s cost and reliability as optimization object and restriction

a network topology optimization model is established.Taking recursive decomposition algorithm as a tool for system reliability analysis

the genetic algorithms combined with sensitivity analysis of system’s elements is applied to solve above optimization problem.An example is given at the end of the paper

and the results show that the suggested algorithm represents certain global searching ability.We can find that networks’ topology positively correlated with system’s reliability capability.