Optimal Spares Distribution Using Probabilistic Evolutionary Algorithms

  • Nouf Alkaabi

Student thesis: Master's Thesis


In the presented thesis work, we are motivated by a real-world spare part allocation problem trying to enhance a tool that is used for automating the process of spare parts allocation in EBTIC's partner telecom. It is a large telecom organization with many inventories and many spare parts scattered across the UK. Their main goal is to find the best distribution of spare parts that will maximize the parts' coverage and minimize the traveling time for supplying a part to a specific location if needed. It is a large-scale real-world problem that can be solved as an optimization problem. In this work, we implement two different evolutionary algorithms (EA) to find a distribution schedule for the spare part movement that will reduce the travel required to move the spare parts, keep the parts closer to the sites where similar parts are deployed, and keep the parts closer to sites where faults are likely to happen. Our work can be divided into two main parts. The first part is implementing a solution based on one of the popular EAs called Random Key-based Estimation of Distribution Algorithm (RKEDA) and compare its performance to a previous solution that implemented a permutation-based Genetic Algorithm (GA). To enhance the implemented RKEDA performance, further research was undertaken in the second part of our work to modify the original version of RKEDA and develop a new version of RKEDA that should enhance the process of reducing the total cost function and converging to the best solution. The algorithms are written in Java and tested using eclipse-workspace.
Date of AwardMay 2021
Original languageAmerican English


  • Genetic Algorithm
  • RK-EDA
  • Resource management
  • spare parts
  • telecom industry
  • inventories.

Cite this