Solving a Redundancy Allocation Problem by a Hybrid Multi-objective Imperialist Competitive Algorithm

تخصيص اجزا مازاد يكي از مسايل معروف NP-Hard مي باشد كه شامل انتخاب اجزا و سطوح افزونگي براي بيشينه نمودن قابليت اطمينان تحت محدوديت هاي مختلف سيستم است. در اكثر طراحي ها، به دليل وجود توابع هدف چندگانه متضاد، سهم بندي قابليت اطمينان سخت و دشوار مي شود. در اين مدل، سه هدف شامل بيشينه سازي قابليت اطمينان و كمينه سازي حجم و هزينه مد نظر قرار مي گيرد كه براي حل آن، يك الگوريتم تلفيقي چندهدفه جديد برپايه الگوريتم هاي ژنتيك و رقابت استعماري براي اولين بار در مسايل تخصيص اجزا مازاد پيشنهاد مي شود. علاوه بر اين از رويه شناسي سطح پاسخ براي تنظيم عملگرهاي الگوريتم پيشنهادي استفاده مي شود. الگوريتم پيشنهادي در مقايسه با دو الگوريتم NSGA-II و PAES از كارايي بالاتري برخوردار است

A redundancy allocation problem (RAP) is a well-known NP-hard problem including the selection of elements and redundancy levels to maximize the system reliability under various system-level constraints. In many practical designing situations, reliability apportionment turns to be complicated because of the presence of several conflicting objectives that cannot be combined into a single-objective function. As in telecommunications, manufacturing and power systems are becoming more and more complex. It is becoming increasingly important to develop efficient solutions to the RAP, while requiring short developments schedules and very high reliability. In this paper, a new hybrid multi-objective imperialist competition algorithm (HMOICA), based on imperialist competitive algorithm (ICA) and genetic algorithm (GA) is proposed in multi-objective redundancy allocation problems. In the multi-objective formulation, system reliability is maximized in which cost and volume of the system are minimized simultaneously. In addition, a response surface methodology (RSM) is employed for parameter tuning of ICA. The proposed HMOICA has also been validated by some examples with analytical solutions. It shows its superior performance compared to a non-dominated sorting genetic algorithm (NSGA-II) and Pareto archive evolution strategy algorithm (PAES).