Perturbation Strategies for the Component Assignment Problem in Three-Dimensional Consecutive-k Systems  
Author Taishin Nakamura

 

Co-Author(s) Tomoaki Akiba; Koji Shingyochi

 

Abstract This paper addresses the component assignment problem in linear connected-(1,1,2)!-out-of-(n1,n2,n3):F systems, where system failure occurs if and only if a pair of adjacent failed components is arranged along any of the three coordinate axes. We propose an iterated local search algorithm incorporating five different perturbation methods specifically designed for this system. Computational experiments on various system sizes reveal that the effectiveness of these perturbation methods depends on the system’s size parameters. Notably, the parity (even or odd) of these parameters influences the optimal component arrangement, as certain advantageous patterns are more easily achieved in systems with odd parameters. The proposed perturbation strategies can also serve as mutation operators in genetic algorithms and other metaheuristics.

 

Keywords Three-dimensional Consecutive-k System, Component Assignment Problem, Iterated Local Search, Optimization
   
    Article #:  RQD2025-188
 

Proceedings of 30th ISSAT International Conference on Reliability & Quality in Design
August 6-8, 2025