| 1. |
- Bagloee, Saeed Asadi, et al.
(author)
-
A hybrid branch-and-bound and Benders decomposition algorithm for the network design problem
- 2017
-
In: Computer-Aided Civil and Infrastructure Engineering. - : Wiley. - 1093-9687 .- 1467-8667. ; 32:4, s. 319-343
-
Journal article (peer-reviewed)abstract
- Given a set of candidate road projects associated with costs, finding the best subset with respect to a limited budget is known as the network design problem (NDP). The NDP is often cast in a bilevel programming problem which is known to be NP-hard. In this study, we tackle a special case of the NDP where the decision variables are integers. A variety of exact solutions has been proposed for the discrete NDP, but due to the combinatorial complexity, the literature has yet to address the problem for large-size networks, and accounting for the multimodal and multiclass traffic flows. To this end, the bilevel problem is solved by branch-and-bound. At each node of the search tree, a valid lower bound based on system optimal (SO) traffic flow is calculated. The SO traffic flow is formulated as a mixed integer, non-linear programming (MINLP) problem for which the Benders decomposition method is used. The algorithm is coded on a hybrid and synchronized platform consisting of MATLAB (optimization engine), EMME 3 (transport planning module), MS Access (database), and MS Excel (user interface). The proposed methodology is applied to three examples including Gao's network, the Sioux-Falls network, and a real size network representing the city of Winnipeg, Canada. Numerical tests on the network of Winnipeg at various budget levels have shown promising results.
|
|
| 2. |
|
|
| 3. |
- Bagloee, Saeed Asadi, et al.
(author)
-
A hybrid machine-learning and optimization method to solve bi-level problems
- 2018
-
In: Expert Systems with Applications. - : Elsevier BV. - 0957-4174. ; 95, s. 142-152
-
Journal article (peer-reviewed)abstract
- © 2017 Elsevier Ltd Bi-level optimization has widespread applications in many disciplines including management, economy, energy, and transportation. Because it is by nature a NP-hard problem, finding an efficient and reliable solution method tailored to large sized cases of specific types is of the highest importance. To this end, we develop a hybrid method based on machine-learning and optimization. For numerical tests, we set up a highly challenging case: a nonlinear discrete bi-level problem with equilibrium constraints in transportation science, known as the discrete network design problem. The hybrid method transforms the original problem to an integer linear programing problem based on a supervised learning technique and a tractable nonlinear problem. This methodology is tested using a real dataset in which the results are found to be highly promising. For the machine learning tasks we employ MATLAB and to solve the optimization problems, we use GAMS (with CPLEX solver).
|
|
| 4. |
- Bagloee, Saeed Asadi, et al.
(author)
-
Optimization for Roads' Construction: Selection, Prioritization, and Scheduling
- 2018
-
In: Computer-Aided Civil and Infrastructure Engineering. - : Wiley. - 1093-9687 .- 1467-8667. ; 33:10, s. 833-848
-
Journal article (peer-reviewed)abstract
- Computer-Aided Civil and Infrastructure Engineering Limited resources (budget, labor, machinery) have a significant toll on the roads' construction. The question of interest is: given variations of resources over a lengthy construction time, what would be the best construction scheduling plan, or how to optimize the Gantt chart while considering two highly challenging features (1) prerequisite conditions and (2) the interdependency of the benefit of the projects’ completions. We formulate it as a bilevel problem where the objective function is to minimize generalized costs and the lower level accounts for the drivers’ route choice. We employ a solution algorithm based on a supervised learning technique (a linear regression model of machine-learning) and an integer programming problem and it is applied to the datasets of Winnipeg and Chicago. The regression model was found to be a tight approximation which resulted in an efficient algorithm (the CPU time is almost a linear function of the number of iterations). Moreover, the proposed methodology can render promising results (at least locally optimal solutions). This article is the first to formulate the Gantt chart using linear binary constraints and optimize it tailored to real-life case studies.
|
|
