Dissertation
Multi-objective Evolutionary Algorithms for Optimal Scheduling
Multi-objective optimization is an effective technique for finding optimal solutions that balance several conflicting objectives. It has been applied in many fields of our world, because practical problems usually have more than one desired goal. For example, developing a new vehicle component might involve minimizing weight while maximizing strength; choosing a portfolio might involve maximizing the expected return while minimizing the risk.
- Author
- Wang, Y.
- Date
- 20 January 2022
- Links
- Thesis in Leiden Repository
The multi-objective optimization problems solved in the thesis originated from the CIMPLO (Cross-Industry Predictive Maintenance Optimization Platform) project. In the CIMPLO project, given a vehicle fleet, each vehicle comprises a set of components, and each component can be maintained in a workshop among given workshops with varying processing times and costs. The goal is to fid the best maintenance order, location and time for each component, i.e., to optimize the maintenance schedule of the vehicle fleet. The maintenance schedule of the vehicle fleet is optimized to bring business advantages to industries, such as, to reduce maintenance time, increase safety, and decrease repair expenses. This problem is strongly NP-hard since the flexible job shop scheduling problem (FJSP) has been proven to be a strongly NP-hard problem and the vehicle fleet maintenance scheduling optimization (VFMSO) problem can be seen as an extension of the FJSP: the VFMSO provides a non-specific operation sequence and involves the processing costs of the operation on machines besides the processing times as in the FJSP. Therefore, evolutionary algorithms (EAs) have been chosen to solve our real-world application problem because they have proven to be a particularly suitable metaheuristic method to solve multi-objective optimization problems.
The rest of this summary can be read in the Repository. See link on top of this page.