Use this URL to cite or link to this record in EThOS:
Title: A decomposition-based ecosystem-inspired approach for solving real-world logistics problems
Author: Adham, Manal Tarek
ISNI:       0000 0004 7660 7503
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Restricted access.
Access from Institution:
Nature exhibits fascinating phenomena which have been used as a source of inspiration for combinatorial optimisation. Many widely accepted nature-inspired approaches focus on specific phenomena like the evolution of biological processes and the behaviour of social groups. This thesis adopts a holistic, bottom-up approach inspired by the synergy between species and their environments. An ecosystem-inspired, decomposition-based approach is proposed: the Artificial Ecosystem Algorithm. Focus is placed on exploring the space of problem subdivisions as opposed to solution spaces, thereby allowing us to exploit the scaling enabled through decomposition. The algorithm was applied to both static and dynamic settings of the travelling salesman problem. The results indicate that the Artificial Ecosystem Algorithm was able to find solutions which are competitive with some present in the literature. Furthermore, they suggest that the algorithm was able to effectively retain solution fragments for dynamic problems. The algorithm was extended and applied to the rebalancing of bikes in Transport for London's public cycle scheme, where it was shown to outperform historical performance levels. It was then further developed and applied to ArcelorMittal's multi-line steel scheduling problem, where it was compared against the current solution implemented and against existing solutions of various types. In this way the Artificial Ecosystem Algorithm is applied to gradually more difficult logistics problems: from static to dynamic, to multi-objective, and finally to constrained. Different features from natural ecosystems are incorporated into the algorithm to improve its ability to solve each problem.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available