Column Generation (CG) is an effective method for solving large-scale optimization problems. CG starts by a subproblem with subset of columns (i.e., variables) and gradually includes new that can improve the solution current subproblem. The are generated as needed repeatedly pricing problem, which often NP-hard bottleneck approach. To tackle this, we propose Machine-Learning-based Pricing Heuristic (MLPH) generate many high-quality efficiently. In each iteration CG, our MLPH leverages ML...

This paper proposes a novel primal heuristic for Mixed Integer Programs, by employing machine learning techniques. Programming is general technique formulating combinatorial optimization problems. Inside solver, heuristics play critical role in finding good feasible solutions that enable one to tighten the duality gap from outset of Branch-and-Bound algorithm (B&B), greatly improving its performance pruning B&B tree aggressively. In this paper, we investigate whether effective can be...

In this paper, we would like to demonstrate an intelligent traffic analytics system called T4, which enables over real-time and historical trajectories from vehicles. At the front end, visualize current flow result of different types queries, as well histograms lights. back T4 is able support multiple common queries trajectories, with compact storage, efficient index fast pruning algorithms. The output those can be used for further monitoring purposes. Moreover, train deep models prediction...

Column Generation (CG) is an effective method for solving large-scale optimization problems. CG starts by a sub-problem with subset of columns (i.e., variables) and gradually includes new that can improve the solution current subproblem. The are generated as needed repeatedly pricing problem, which often NP-hard bottleneck approach. To tackle this, we propose Machine-Learning-based Pricing Heuristic (MLPH)that generate many high-quality efficiently. In each iteration CG, our MLPH leverages...

Column generation (CG) is a well-established method for solving large-scale linear programs. It involves iteratively optimizing subproblem containing subset of columns and using its dual solution to generate new with negative reduced costs. This process continues until the values converge optimal original problem. A natural phenomenon in CG heavy oscillation during iterations, which can lead substantial slowdown convergence rate. Stabilization techniques are devised accelerate by information...

Column generation (CG) is a powerful technique for solving optimization problems that involve large number of variables or columns. This begins by smaller problem with subset columns and gradually generates additional as needed. However, the often requires difficult subproblems repeatedly, which can be bottleneck CG. To address this challenge, we propose novel method called machine learning enhanced ant colony (MLACO), to efficiently generate multiple high-quality from subproblem....

This paper introduces an enhanced meta-heuristic (ML-ACO) that combines machine learning (ML) and ant colony optimization (ACO) to solve combinatorial problems. To illustrate the underlying mechanism of our ML-ACO algorithm, we start by describing a test problem, orienteering problem. In this objective is find route visits subset vertices in graph within time budget maximize collected score. first phase ML model trained using set small problem instances where optimal solution known....

This paper aims to predict optimal solutions for combinatorial optimization problems (COPs) via machine learning (ML). To find high-quality efficiently, existing work uses a ML prediction of the solution guide heuristic search, where model is trained offline under supervision solved problem instances with known solutions. sufficient accuracy, it critical provide adequate features that can effectively characterize decision variables. However, acquiring such challenging due high complexity...

This paper proposes a novel primal heuristic for Mixed Integer Programs, by employing machine learning techniques. Programming is general technique formulating combinatorial optimization problems. Inside solver, heuristics play critical role in finding good feasible solutions that enable one to tighten the duality gap from outset of Branch-and-Bound algorithm (B&B), greatly improving its performance pruning B&B tree aggressively. In this paper, we investigate whether effective can be...