Ant colony optimization approach to heterogeneous redundancy in multi-state systems with multi-state components

• Published in: 2009 8th International Conference on Reliability, Maintainability and Safety pp. 116 - 121
• Main Authors: Sharma, V.K., Agarwal, M.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2009
• Subjects: Algorithm design and analysis; Ant colony optimization; Cost function; Design optimization; heterogeneous redundancy; multi-state components; multi-state system; Plastics; Power generation; Production management; Redundancy; Reliability; series-parallel system; Traveling salesman problems
• DOI: 10.1109/ICRMS.2009.5270227
• ISBN: 1424449030; 9781424449033

An algorithm based on ant colony optimization (ACO) has been developed and employed to address the problem of optimum redundancy allocation in series-parallel multi-state systems (MSS) consisting of multi-state components. The objective is to obtain a minimum cost configuration of the system that satisfies the given reliability and weight constraints and the consumer load demand occurring in different operating time intervals. The demand distribution is presented as a piecewise cumulative load curve. The multi-state components are chosen from a list of products available in the market and have their characteristic feeding capacity, reliability, weight and cost. The capacity of the system, thus, strongly depends upon the selection of constituent components. The algorithm allocates heterogeneous redundancy i.e. non-identical components (of maximum two types) are allowed in each subsystem The search of optimal system structure in the ACO algorithm presented in the paper implements a multinomial probability based method to compute exact system reliability index. A penalty function is coupled to handle the constraints and restrict the search near the feasible region. The algorithm is very easy to apply and still obtains very good solutions with promising time efficiency. Two illustrative examples are given to validate the algorithm and to demonstrate its performance.