Optimization of systems reliability by metaheuristic approach

The application of metaheuristic approaches in addressing the reliability of systems through optimization is of greater interest to researchers and designers in recent years. Reliability optimization has become an essential part of the design and operation of largescale manufacturing systems. This thesis addresses the optimization of system-reliability for series–parallel systems to solve redundant, continuous, and combinatorial optimization problems in reliability engineering by using metaheuristic approaches (MAs). The problem is to select the best redundancy strategy, component, and redundancy level for each subsystem to maximize the system reliability under system-level constraints. This type of problem involves the selection of components with multiple choices and redundancy levels that yield the maximum benefits, and it is subject to the cost and weight constraints at the system level. These are very common and realistic problems faced in the conceptual design of numerous engineering systems. The development of efficient solutions to these problems is becoming progressively important because mechanical systems are becoming increasingly complex, while development plans are decreasing in size and reliability requirements are rapidly changing and becoming increasingly difficult to adhere to. An optimal design solution can be obtained very frequently and more quickly by using genetic algorithm redundancy allocation problems (GARAPs). In general, redundancy allocation problems (RAPs) are difficult to solve for real cases, especially in large-scale situations. In this study, the reliability optimization of a series–parallel by using a genetic algorithm (GA) and statistical analysis is considered. The approach discussed herein can be applied to address the challenges in system reliability that includes redundant numbers of carefully chosen modules, overall cost, and overall weight. Most related studies have focused only on the single-objective optimization of RAP. Multiobjective optimization has not yet attracted much attention. This research project examines the multiobjective situation by focusing on multiobjective formulation, which is useful in maximizing system reliability while simultaneously minimizing system cost and weight to solve the RAP. The present study applies a methodology for optimizing the reliability of a series–parallel system based on multiobjective optimization and multistate reliability by using a hybrid GA and a fuzzy function. The study aims to determine the strategy for selecting the degree of redundancy for every subsystem to exploit the general system reliability depending on the overall cost and weight limitations. In addition, the outcomes of the case study for optimizing the reliability of the series–parallel system are presented, and the relationships with previously investigated phenomena are presented to determine the performance of the GA under review. Furthermore, this study established a new metaheuristic-based technique for resolving multiobjective optimization challenges, such as the common reliability redundancy allocation problem. Additionally, a new simulation process was developed to generate practical tools for designing reliable series–parallel systems. Hence, metaheuristic methods were applied for solving such difficult and complex problems. In addition, metaheuristics provide a useful compromise between the amount of computation time required and the quality of the approximated solution space. The industrial challenges include the maximization of system reliability subject to limited system cost and weight, minimization of system weight subject to limited system cost and the system reliability requirements and increasing of quality components through optimization and system reliability. Furthermore, a real-life situation research on security control of a gas turbine in the overspeed state was explored in this study with the aim of verifying the proposed algorithm from the context of system optimization