-

Ant Colony System (ACS) is defined as a metaheuristic acting according to the behavior of natural ants. In this article ACS is used for reliability optimization of a series system under multiple- choice for the selection of the type of technology in each subsystem such that optimization of the reliability of the whole system can be achieved with regard to budget constraint. This problem is NP-hard and it could be modeled as zero-one nonlinear programming. In this paper we propose an ACS algorithm based on the artificial ant behavior which responding to the communication with previous ants and with regard to the objective function. The constructed solutions are not guaranteed to be feasible; therefore using a convenient mechanism an infeasible solution is replaced by feasible one and then they are improved by a local search. We have compared the proposed method with an Ant System (AS) algorithm. Computational results have shown that the proposed ACS has better performance in producing solutions with higher quality in particular for problems with large number of subsystems.