MARKOV-BASED AVAILABILITY MODELLING AND OPTIMIZATION USING SPIDER MONKEY OPTIMIZATION

Abstract

This paper presents the availability optimization of a screw manufacturing plant using the Spider Monkey Optimization (SMO) algorithm. The plant consists of four major subsystems, two of which operate with cold standby redundancy. Assuming exponential failure and repair distributions, the system is modelled using a probabilistic approach and differential equations are developed through the Markov birth–death process. Steady-state availability is determined using normalization conditions. To enhance system performance, SMO, a swarm-based metaheuristic inspired by the social foraging behaviour of spider monkeys, is employed to maximize availability by optimizing subsystem failure and repair rates within specified bounds. The algorithm simultaneously coordinates twelve decision variables to obtain the optimal parameter combination. The results indicate that SMO significantly improves system availability and provides an efficient and robust alternative to traditional evolutionary optimization techniques for complex industrial systems.