Development of a blended particle swarm optimization to optimize mission design life of a series–parallel reliable system with time dependent component reliabilities in imprecise environments

Abstract

Maximization of system reliability is a topic of much concern in the field of engineering design, telecommunication, aircraft design, plant and factory establishing, producing of sensitive devices, and many production management. The main goal of reliability optimization is to construct highly reliable system or to design maximum reliable system so that the system can run smoothly without any hindrance under some restrictions in the form of constraints for a specified period of time maintaining the safety and efficiency. In this paper, we consider Redundancy Allocation Problem (RAP) involving time dependent component reliabilities for optimizing system reliability of the problem designed. Moreover, we have considered a new objective function as the Mission Design Life (MDL) of reliable system. The MDL is found by integrating the system reliability with respect to time from zero to mission time. This consideration makes the problem more realistic by taking the fact into account that reliability of the system varies with time. Here, we have maximized the Mission Design Life of reliable system with the time dependent component reliabilities by implementing the proposed Blended WQPSO (BWQPSO). The BWQPSO is developed by incorporating the Big-M penalty function in the Weighted Quantum behaved Particle Swarm Optimization and it is found to perform highly satisfactorily to solve the highly nonlinear combinatorial problem so designed as a time dependent redundancy allocation problem. Again, to tackle the uncertain behaviour of controlling parameters, we incorporated the fuzziness and intuitionistic fuzziness to develop the imprecise models. The three types of problems designed here are supplemented by numerical examples. The numerical experiments are carried out in each of the environments and comparative studies are presented. The statistical analysis along with computational time and convergence of the proposed algorithm are also produced.

Redundancy Allocation of Components with Time-Dependent Failure Rates

“…Other aspects come realistically into play in a RAP; these include the objective function, the states of the components and that of the entire system, the type of the components to be allocated to each subsystem, the reparability of the components, and so on. For example, RAPs may be considered with a single objective function, e.g., aiming to maximizing system reliability [1,[3][4][5][6][7][8][9][10][11]27,28,30,31,[35][36][37][39][40][41]44,[46][47][48][53][54][55], minimizing total system cost [12][13][14][15] or maximizing/minimizing other parameters [16,29]. Alternatively, RAPs may consider multi-objective functions [18][19][20][21][22][23][24][25]32,34,38,42,45,50,…”

“…Regarding the components, they might be represented as binary or multi-state. In the binary state representation, the components can only be totally healthy or completely failed [1,[3][4][5][6][7][8][9][10][11][12][13][17][18][19][20][21][22][23][25][26][27][28][29][30][31][32][33][34][35]40,41,[43][44][45][46][47][48]53,55]; in the multi-state, the components might have other states, intermediate between these two [14][15][16]24,[60][61][62][63][64][65]. The type of the components ...…”

“…Moreover, the components characteristics, including their cost, weight and performance function, might be deterministic or uncertain functions [66]. From another aspect, the components can be repairable [14,15,20,24,26,30,56] or non-repairable [1,[3][4][5][6][7]9,11,13,[16][17][18][19]21,22,[27][28][29]31,[35][36][37]40,41,[43][44][45][46][47][48][53][54][55]. In this regard, a comprehensive categorization of reliability optimization problems is provided in [67].…”