Use this URL to cite or link to this record in EThOS:
Title: Evolutionary algorithms for scheduling operations
Author: Khmeleva, Elena
ISNI:       0000 0004 6348 4006
Awarding Body: Sheffield Hallam University
Current Institution: Sheffield Hallam University
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
While business process automation is proliferating through industries and processes, operations such as job and crew scheduling are still performed manually in the majority of workplaces. The linear programming techniques are not capable of automated production of a job or crew schedule within a reasonable computation time due to the massive sizes of real-life scheduling problems. For this reason, AI solutions are becoming increasingly popular, specifically Evolutionary Algorithms (EAs). However, there are three key limitations of previous studies researching application of EAs for the solution of the scheduling problems. First of all, there is no justification for the selection of a particular genetic operator and conclusion about their effectiveness. Secondly, the practical efficiency of such algorithms is unknown due to the lack of comparison with manually produced schedules. Finally, the implications of real-life implementation of the algorithm are rarely considered. This research aims at addressing all three limitations. Collaborations with DBSchenker,the rail freight carrier, and Garnett-Dickinson, the printing company,have been established. Multi-disciplinary research methods including document analysis, focus group evaluations, and interviews with managers from different levels have been carried out. A standard EA has been enhanced with developed within research intelligent operators to efficiently solve the problems. Assessment of the developed algorithm in the context of real life crew scheduling problem showed that the automated schedule outperformed the manual one by 3.7% in terms of its operating efficiency. In addition, the automatically produced schedule required less staff to complete all the jobs and might provide an additional revenue opportunity of £500 000. The research has also revealed a positive attitude expressed by the operational and IT managers towards the developed system. Investment analysis demonstrated a 41% return rate on investment in the automated scheduling system, while the strategic analysis suggests that this system can enable attainment of strategic priorities. The end users of the system, on the other hand, expressed some degree of scepticism and would prefer manual methods.
Supervisor: Shahidan, Malihe Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available