Abstract:The design or the optimization of transport systems is a difficult task. This is especially true in the case of the introduction of new transport modes in an existing system. The main reason is, that even small additions and changes result in the emergence of new travel patterns, likely resulting in an adaptation of the travel behavior of multiple other agents in the system. Here we consider the optimization of future Urban Air Mobility services under consideration of effects induced by the new mode to an existing system. We tackle this problem through a bi-level network design approach, in which the discrete decisions of the network design planner are optimized based on the evaluated dynamic demand of the user's mode choices. We solve the activity-based network design problem (AB-NDP) using a Genetic Algorithm on a multi-objective optimization problem while evaluating the dynamic demand with the large-scale Multi-Agent Transport Simulation (MATSim) framework. The proposed bi-level approach is compared against the results of a coverage approach using a static demand method. The bi-level study shows better results for expected UAM demand and total travel time savings across the transportation system. Due to its generic character, the demonstrated utilization of a bi-level method is applicable to other mobility service design questions and to other regions.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to optimize the layout design of UAM vertiports when introducing a new traffic mode - Urban Air Mobility (UAM) service - into the existing traffic system. Specifically, the researchers focus on how to optimize the location allocation of UAM vertiports through a two - layer network design method to adapt to the new travel patterns caused by the new traffic mode and consider the overall impact of this change on the entire traffic system, such as the savings in total travel time.
### Background and Problem Description of the Paper
With the acceleration of the urbanization process, the problem of urban traffic congestion is becoming increasingly serious, and traditional ground transportation methods are difficult to meet the growing travel demands. Therefore, urban air mobility has been proposed as an emerging transportation mode, aiming to provide fast and efficient point - to - point transportation services by using electric vertical take - off and landing aircraft (eVTOLs). However, the successful implementation of UAM services depends not only on technological progress but also on reasonable infrastructure planning, especially the rational layout of UAM vertiports.
### Research Objectives
The main objective of the paper is to develop a two - layer network design framework to solve the optimal location allocation problem of UAM vertiports. This framework needs to be able to:
1. **Consider Dynamic Demands**: Through activity - based transport simulation, evaluate the impact of different vertiport layout schemes on users' travel choices and how these choices further affect the operating state of the entire traffic system.
2. **Multi - objective Optimization**: Optimize multiple objectives simultaneously, such as maximizing the demand for UAM services and minimizing the number of required vertiports.
3. **Comprehensively Consider Multiple Factors**: Not only focus on users who directly use UAM services, but also consider the reactions of users of other traffic modes and the impact of these reactions on traffic flow, congestion levels, etc.
### Methods and Technologies
To achieve the above - mentioned objectives, the researchers adopted the following methods and technologies:
- **Two - layer Optimization Framework**: The outer layer uses the Genetic Algorithm (GA) for network design optimization, and the inner layer uses the Multi - Agent Transport Simulation (MATSIM) framework to evaluate dynamic demands.
- **Multi - objective Optimization**: Use the Non - Dominated Sorting Genetic Algorithm (NSGA - II) for multi - objective optimization, with objectives including maximizing UAM demand and minimizing the number of vertiports.
- **Activity - based Transport Simulation**: Through MATSIM, simulate the activity chains of a large number of agents within 24 hours to evaluate the impact of different network layout schemes on the traffic system.
### Experiments and Results
The researchers conducted experiments on an open - source traffic simulation platform, using the traffic data of Corsica as a case. By comparing the two - layer optimization method (AB - NDP) with the traditional static coverage method (Heuristic Coverage Method, HCM), the results show that:
- **UAM Demand**: The AB - NDP method increased the UAM demand by 19% compared to the HCM method.
- **Total Travel Time Savings**: The AB - NDP method increased the total travel time savings by approximately 1.62% (from 5.65% to 7.27%) compared to the HCM method.
### Conclusions
The two - layer optimization framework AB - NDP proposed in the paper performs excellently in optimizing the layout of UAM vertiports. It not only increases the demand for UAM services but also significantly improves the efficiency of the entire traffic system. Although this method increases complexity, it can consider various factors more comprehensively and provides strong support for the future planning of urban air mobility.