FAULT TOLERANCE MAXIMIZATION OF HARDWARE NETWORKS UNDER BUDGET CONSTRAINT USING CORONA VIRUS OPTIMIZATION ALGORITHM

Abstract

The paper intends to maximize faults tolerance in hardware networks design. A series parallel design is proposed where each component kind like a switch version or a server version is characterized by its price and its reliability. The redundancy optimization problem (ROP) or system design optimization consists of choosing an adequate number of redundant components. In order to maximize the reliability under available cost, a metaheuristic which is the corona virus optimization algorithm (CVOA) is adapted to the design problem. The CVOA is inspired by the propagation of the virus when each infected, recovered, or dead patient represents a solution. The population size and the time can be unknown at first. The hill climbing with neighborhood function is added to the original algorithm to generate new infections. Computational Experiments are performed on a heterogeneous network with 10 subsystems using CICSO components and on a real shipping company architecture with 6 subsystems. They confirm the feasibility of the adapted algorithm to maximize the reliability of the networks without outcoming the allowed budget and reveal its efficiency compared to other swarm metaheuristics like interior search algorithm, bat algorithm and bat algorithm with generalized flight.