University of Twente Student Theses


Improving routes of a repair company

Engelen, Julian van (2022) Improving routes of a repair company.

[img] PDF
Abstract:Hamer is a technical installation company, which provides services and installations in the utility, industry, automotive and fuel sectors. Numerous Hamer departments are located throughout the Netherlands and Belgium, with the headquarters being located in Apeldoorn. The core activity of Hamer consists of the installation of technical machinery and the maintenance of these devices. This research focuses on the fuel department, which mainly repairs malfunctions spread throughout the Netherlands. Hamer is constantly seeking for higher efficiency, but currently, a mechanic can handle 20 malfunctions in a week on average, while their norm is 21. After the current situation had been investigated, it was decided that the routes of the mechanics could provide the necessary improvements. Thereafter, in-depth observation and literature research has been conducted to learn more about the problem. The problem is a Vehicle Routing Problem (VRP) and after research it is classified as an adapted Multi-Depot Vehicle Routing Problem with Time Windows (MDVRPTW). Currently, the routes are made manually a day prior without any software assistance. Hamer has to handle various malfunctions per day, but they are limited by the current capacity of employees. Next to that, malfunctions have a difference in priority due to the Service Level Agreements (SLAs) Hamer has with their clients. Therefore the routes need to be based on the distance and priority of jobs. The VRP is an NP-hard problem (nondeterministic polynomial time), meaning that computing the optimal solution could take non-affordable time. Therefore certain heuristic algorithms were chosen, which gave a sufficient solution in reasonable time. The heuristics take into account the distance and priority of associated jobs. For creating an initial solution, an algorithm is made using an adapted Push Forward Insertion Heuristic (PFIH). In order to optimize the initial solution, the following improvement heuristics were used: an adapted 2-interchange as inter-route operator, and 2-opt as intra-route operator. The Route Planning Software (RPS) decision support tool for Hamer was made using Python and Excel. A single-day approach is utilized to determine the routes for the following day. This was done because of additional malfunctions that often occur throughout the week, thus making it difficult to plan far ahead of time. The software produces routes for Hamer in relatively short computational time, allowing Hamer to use it to find efficient routes quickly. Using the self-made RPS decision support tool with the explained heuristics, various experiments were conducted. The priority of locations can be set to Hamer’s desires, so that Hamer can improve certain Key Performance Indicators (KPIs). There are different experiments conducted to evaluate the algorithms and parameters. In addition, an experiment was carried out to compare the outcomes of the real situation with the results recommended by the RPS. The outcomes were attained at a time when there were, on average, 5.6 mechanics available. The results showed a predicted change in the average distance traveled from 53.0 km per failure to 48.1 km. Additionally, the findings demonstrated that 1.25 more jobs could be performed on average by each mechanic in a week while requiring them to work 13.4 hours less overall. In total the distance driven weekly was reduced from 6044 km to 5821, while handling 121 jobs instead of 114. The primary proposal for Hamer after this research is to conduct further research concerning RPS. Hamer has many additional departments where the implementation of RPS could increase productivity. Due to time constraints, the problem could not be fully resolved and implemented into Hamer’s departments. For further implementation, Hamer could adapt the software to attain their wishes. Firstly, the provided solution is still a single-period approach, whereas Hamer faces a multi-period problem. Additionally, the distance and driving time are now based on the orthodromic distance and average driving speeds, which could be modified using Add-ins. Finally, the emphasis was solely on mechanics, which are capable of resolving any malfunctions in the fuel department. However, there is a minor amount of mechanics that do not possess the required skill set to solve a certain problem. Certain adaptations to the RPS could ensure that all mechanics are included with their personalized skills taken into account. Thus, excluding certain mechanics from malfunctions they are unable to resolve. Further research in these areas would result in a better reflection of reality and thus more accurate route optimization.
Item Type:Essay (Bachelor)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Industrial Engineering and Management BSc (56994)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page