Combining a maintenance center M/M/c/m queue into the economic production quantity model with stochastic machine breakdowns and repair

  • Authors:
  • Kamyar Sabri-Laghaie;Mostafa Mansouri;Arash Motaghedi-Larijani;Gholamreza Jalali-Naini

  • Affiliations:
  • Industrial Engineering Department, Iran University of Science & Technology, Farjam Street, Narmak, Tehran, Iran;Industrial Engineering Department, Iran University of Science & Technology, Farjam Street, Narmak, Tehran, Iran;Industrial Engineering Department, Iran University of Science & Technology, Farjam Street, Narmak, Tehran, Iran;Industrial Engineering Department, Iran University of Science & Technology, Farjam Street, Narmak, Tehran, Iran

  • Venue:
  • Computers and Industrial Engineering
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

This article considers a production system in which ''m'' identical machines produce nonidentical products at production rates. Products made by each machine are on the other hand consumed at a specific demand rate. Machines may be affected by unwanted breakdowns. Machines break down according to a Poisson distribution with equal rates and the failed machines are sent to maintenance center for repair which is consisted of ''c'' servers or servicemen. However, Number of machines is greater than number of servicemen (mc). Hence, if the number of failed machines is greater than that of servicemen, the machines have to be put in a queue. Machines are put in one queue with the order of queue being FCFS. The queue has a typical M/M/c/m system. If machines are broken down during production, shortage will occur. This problem has been considered to obtain a single production cycle for the machines and an optimum number of the servers such that costs are minimized. For this purpose, distribution of waiting time for machines in repair center is calculated and a cost function is formed. Steepest descent and direct search methods are applied in this work to obtain optimal production cycle and maintenance servers, respectively. The proposed methods are studied using a comprehensive example.