This paper develops an approach for automation-assisted design of traffic flow management strategies using multiobjective genetic algorithms. In situations where the imbalance of demand and capacity cannot be managed with tactical measures, traffic management initiatives are often imposed; however, the definition of these initiatives, including a number of parameters required for implementation, are determined based on experience. For automation to assist in this process, a method is needed to identify a small number of solutions that embody the best strategies across a variety of metrics. This paper proposes an approach that uses multiobjective genetic algorithms to readily identify the Pareto set of solutions and further cluster these solutions based on the structure of the strategies as well as their performance. Results verify the performance, accuracy, and scalability of the multiobjective genetic algorithm for identifying the Pareto set of solutions by evaluating only a minute fraction of the design space. Furthermore, by employing the clustering and solution selection procedure developed in this paper, the goal of identifying a small number of distinct solution alternatives for further evaluation is achieved.
Designing Traffic Flow Management Strategies Using Multiobjective Genetic Algorithms
Journal of Guidance, Control, and Dynamics ; 38 , 10 ; 1922-1934
2015-02-27
13 pages
Article (Journal)
Electronic Resource
English
Designing Traffic Flow Management Strategies Using Multiobjective Genetic Algorithms
Online Contents | 2015
|Constraint handling in multiobjective genetic algorithms
AIAA | 2000
|Transportation Research Record | 1999
|