This repository assembles source codes and materials of a model developed as part of a master thesis project by H. Miyoshi at Delft University of Technology, MSc Civil Engineering, Traffic and Transport Engineering track. As the circumstances of urban transportation have aggravated due to rapid urbanisation, combining passenger mobility and parcels in the urban transportation system has been proposed as an innovative solution to address these issues. To investigate the impact of such urban transportation systems and the possibility of operating them effectively, this project aimed to develop a dynamic model that optimises the operations of waterborne vessels that serve passenger mobility and parcels in the urban environment. A rolling horizon approach was taken to implement the dynamicity in the model, and an insertion heuristic was proposed to solve the problem at each step. Moreover, a non-homogeneous Poisson process was applied to create individual requests from passengers to incorporate the stochasticity of demand for the service. Through a series of experiments, we demonstrated the capability of the developed insertion heuristic and the improvement of transportation performance by providing mixed service.
- Development of a mathematical model for the dynamic pickup and delivery problem for heterogeneous services using a rolling horizon approach.
- Development of an insertion heuristic to solve the routing of vessels at each time step.
- Evaluation of the insertion heuristic through a series of static experiments between the proposed heuristic and the exact method.
- Development of a stochastic approach to create individual requests throughout the time horizon.
- Comparison of transport performance between mixed-purpose and fixed-purpose vessels through KPI analysis under different demand scenarios.
The repository consists of multiple folders.
- Papers: Contains the full report about this project.
- Presentations: Contains the PowerPoint file which summarises the project.
- Src: Contains the source codes developed in this project and sample notebooks to exhibit the model yourself.
- data: Contains sample parameter data files to create requests in the sample notebook.
- Config: Contains a .txt file listing the necessary packages to run the source codes.
- The full report can be found in "Papers" folder or https://repository.tudelft.nl/record/uuid:750990c5-2911-406c-9db9-89904b1520c7
- H.Miyoshi: Author of the report and developer of source codes.
- Dr. Y.Zhang: Modelling supervision.
- Dr. S.S.Azadeh: Main Supervisor of the project.
- Prof. Dr. O.Cats: Supervisor of the project.
- J.Cheng: Github repository setup support.
- L.Gomes: Github repository setup support.