Robust Design of Transportation Networks in a Scenario-Based Demand Condition with Genetic and Ant Colony Algorithms

Transportation network design is one of the most important issues in transportation engineering, which is complicated by uncertainty. Uncertainty in demand is one of the most common types of uncertainty in traffic networks, which comes from various sources such as difficulties in forecasting of trip productions and attractions. In this paper, a method based on robust optimization for discrete network design in scenario-based variable demand conditions is used, in which the level of imposed uncertainty of that problem will depend on designer choice. For this reason, the objective functions of weighted mean and variance total travel time in several demand scenarios are implemented; and for solving that problem, genetic and ant colony algorithms are used. To examine the proposed model, two case studies are applied to an ordinary network. In the first one, the demand pattern in hypothetical scenarios has a high disparity, and the variance of total travel time is more important in the final design than the mean. In the second case study, the demand pattern has a low disparity, and the weight of the mean and variance is more significant in the design solution. Based on two conditions, the results of design in the condition of uncertainty are different with independent design of each scenario, which emphasizes the use of optimization methods in conditions of uncertainty to solve similar problems. Also, in comparison of genetic and ant colony algorithms in various weights of the objective function, the ant colony algorithm found better solutions with less calculation effort.