Use this URL to cite or link to this record in EThOS:
Title: Applying computational intelligence to a real-world container loading problem in a warehouse environment
Author: Remi-Omosowon, Ayodeji
ISNI:       0000 0004 7228 8875
Awarding Body: Nottingham Trent University
Current Institution: Nottingham Trent University
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
One of the problems presented in the day-to-day running of a warehouse is that of optimally selecting and loading groups of heavy rectangular palletised goods into larger rectangular containers while satisfying a number of practical constraints. The research presented in this thesis was commissioned by the logistics department in NSK Europe Ltd, for the purpose of providing feasible solutions to this problem. The problem is a version of the Container Loading Problem in the literature, and it is an active research area with many practical applications in industry. Most of the advances made in this area focus more on the optimisation of container utility i.e. volume or weight capacity, with very few focusing on the practical feasibility of the loading layout or pattern produced. Much of the work done also addresses only a few practical constraints at a time, leaving out a number of constraints that are of importance in real-world container loading. As this problem is well known to be a combinatorial NPhard problem, the exact mathematical methods that exist for solving it are computationally feasible for only problem instances with small sizes. For these reasons, this thesis investigates the use of computational intelligence techniques for solving and providing near-optimum solutions to this problem while simultaneously satisfying a number of practical constraints that must be considered for the solutions provided to be feasible. In proposing a solution to this problem and dealing with all the constraints considered, an algorithmic framework that decomposes the CLPs into sub-problems is presented. Each subproblem is solved using an appropriate algorithm, and a combination of constraints particular to each problem is satisfied. The resulting hybrid algorithm solves the entire problem as a whole and satisfies all the considered constraints. In order to identify and select feasible container layouts that are practical and easy to load, a measure of disorder, based on the concept of entropy in physics and information theory, is derived. Finally, a novel method of directing a Monte-Carlo tree search process using the derived entropy measure is employed, to generate loading layouts that are comparable to those produced by expert human loaders. In summary, this thesis presents a new approach for dealing with real-world container loading in a warehouse environment, particularly in instances where layout complexity is of major importance; such as the loading of heavy palletised goods using forklift trucks. The approach can be used to deal with a number of relevant practical constraints that need to be satisfied simultaneously, including those encountered when the heavy goods are arranged and physically packed into a container using forklift trucks.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available