Heuristic algorithm for optimal redundancy allocation in complex systems

Abstract

Purpose The purpose of this paper is to present an efficient heuristic algorithm based on the 3-neighborhood approach. In this paper, search is made from sides of both feasible and infeasible regions to find near-optimal solutions. Design/methodology/approach The algorithm performs a series of selection and exchange operations in 3-neighborhood to see whether this exchange yields still an improved feasible solution or converges to a near-optimal solution in which case the algorithm stops. Findings The proposed algorithm has been tested on complex system structures which have been widely used. The results show that this 3-neighborhood approach not only can obtain various known solutions but also is computationally efficient for various complex systems. Research limitations/implications In general, the proposed heuristic is applicable to any coherent system with no restrictions on constraint functions; however, to enforce convergence, inferior solutions might be included only when they are not being too far from the optimum. Practical implications It is observed that the proposed heuristic is reasonably proficient in terms of various measures of performance and computational time. Social implications Reliability optimization is very important in real life systems such as computer and communication systems, telecommunications, automobile, nuclear, defense systems, etc. It is an important issue prior to real life systems design. Originality/value The utilization of 3-neighborhood strategy seems to be encouraging as it efficiently enforces the convergence to a near-optimal solution; indeed, it attains quality solutions in less computational time in comparison to other existing heuristic algorithms.

Minimal path set importance in complex systems

“…Now consider 5 unit Bridge system for nonlinear constraints as in Example 1 of Aggarwal. 30 Table 3 shows the procedure to obtain the optimal solution.…”

“…Agarwal and Aggarwal 16 showed that their algorithms perform better than those in Kohda and Inoue and Shi 18,19 and provide attractive alternatives to that of Kim and Yum. 20 Recently, Aggarwal 30 proposed a heuristic algorithm based on 3-neighborhood, which is an improvement over Agarwal et al 25…”

“…A huge number of techniques, including heuristic and metaheuristic, have been developed for solving redundancy allocation problems for complex systems in the literature. 16,[18][19][20][21][22][23][24][25][26][27][28][29][30] Shi 19 proposed a heuristic algorithm to solve problems of redundancy allocation based on minimal path sets. Here, the feasible solutions obtained in successive iterations are in 1-neighborhood or 2-neighborhoods of each other.…”

“…Agarwal and Aggarwal 16 showed that their algorithms perform better than those in Kohda and Inoue and Shi 18,19 and provide attractive alternatives to that of Kim and Yum. 20 Recently, Aggarwal 30 proposed a heuristic algorithm based on 3-neighborhood, which is an improvement over Agarwal et al 25 For the broadly considered issue of redundancy allocation in complex systems, we introduce a heuristic algorithm P-Alg 2 that does not require the constraints to be separable and monotonically non-decreasing. The algorithm P-Alg 2 requires determining all minimal path sets of the system following an approach of adjusting unit-increment with time.…”