Multiobjective optimization of availability and cost in repairable systems design via genetic algorithms and discrete event simulation

Abstract

This paper attempts to provide a more realistic approach to the characterization of system reliability when handling redundancy allocation problems: it considers repairable series-parallel systems comprised of components subjected to corrective maintenance actions with failure-repair cycles modeled by renewal processes. A multiobjective optimization approach is applied since increasing the number of redundancies not only enlarges system reliability but also its associated costs. Then a multiobjective genetic algorithm is coupled with discrete event simulation and its solutions present the compromise between system reliability and cost. Two examples are provided. In the first one, the proposed algorithm is validated by comparison with results obtained from a system devised as to allow for analytical solutions of the objective functions. The second case analyzes a repairable system subjected to perfect repairs. Results from both examples show that the proposed method can be a valuable tool for the decision maker when choosing the system design.Keywords: availability, redundancy allocation, multiobjective optimization in system design. ResumoEsse artigo utiliza uma abordagem mais realista para a caracterização da confiabilidade de sistemas em problemas de alocação de redundâncias: são considerados sistemas série-paralelo formados por componentes sujeitos a ações de manutenção corretiva com ciclos de falha-reparo modelados por processos de renovação. É aplicada uma abordagem de otimização multiobjetivo, pois o aumento de redundâncias eleva a confiabilidade do sistema e também os seus custos. Assim, um algoritmo genético multiobjetivo é integrado com simulação discreta de eventos e suas soluções apresentam o compromisso entre confiabilidade e custo do sistema. Dois exemplos são fornecidos. No primeiro, o algoritmo proposto é validado através da comparação com resultados obtidos de um sistema criado de forma a permitir soluções analíticas das funções-objetivo. No segundo, analisa-se um sistema reparável sujeito a reparos perfeitos. Os resultados mostram que o método proposto pode ser uma ferramenta valiosa para o decisor no momento da escolha do projeto do sistema.

Modeling and Simulation in Cyber‐Risk

“…The closest article to this one [30] uses only four transition rates in a three-state Markov model, whereas the reviewed MC model by the author uses all six transitions. Other close references deal with different topics; however, none use any simulation techniques [31,32]. Other close references deal with different topics; however, none use any simulation techniques [31,32].…”

Modeling and simulation in engineering

“…In the past, various articles have studied a similar problem. For instance, ‘ A Hybrid Markov system dynamics approach for availability analysis of degraded systems ’ by Rao et al30; similarly, Lins and Droguett study ' Multi‐objective Optimization of Availability and Cost in Repairable Systems Design Via Genetic Algorithms and Discrete Event Simulation 31. ‘ Reliability and Availability Analysis of Three‐state Device Redundant Systems with Human Errors and Common‐cause Failures ’, by Shah and Dhillon studies somewhat similar but still different topics 32.…”

“…However, this review article's simulation approach is even more powerful and flexible as the application can be extrapolated based on identical principles to four or more states, whereas reference by Rao et al30 deals solely with differential equations limited in scope. Others, Lins et al31 and Shah et al,32 are on slightly different but not identical topics; all of which do not employ modeling and simulation techniques or generate closed form statistical probability density function (p.d.f.) expressions and derivations.…”

Two tuned multi-objective meta-heuristic algorithms for solving a fuzzy multi-state redundancy allocation problem under discount strategies

“…Khalili-Damghani et al [5] used a decision support system for solving multi-objective RAPs, where they applied the technique for order performance by similarity to ideal solution (TOPSIS) method to decrease the dimension. Other works on RAP of binary-state systems are due to [6][7][8][9] for repairable and [10][11][12][13][14] for non-repairable components.…”