03:30 PM - 03:55 PM
A GRASP + ILP Based Heuristic for the Capacitated Location-Routing Problem
We present a three-stage heuristic method for the capacitated location-routing problem (CLRP). In a first stage, the CLRP is solved by a GRASP method. In the second stage, the different solutions
provided by the GRASP are embedded into a ILP solver. In the third and last stage, we perform a destroy-and-repair method based on the iterative solution of the same integer program. We present
computational results that show the effectiveness of our approach.
03:55 PM - 04:20 PM
A Piecewise Linearization Approach for a Two-Echelon Inventory Location Problem
This paper considers a model that determines plant and depot locations, shipment levels from plants to depots, safety‐stock levels at depots, and the assignment of clients to depots by minimizing the sum of fixed facility location, transportation, and safety‐stock costs. The model is formulated as a MINLP and linearized using piecewise‐linear functions.
04:20 PM - 04:45 PM
Supermodular Properties in Hub Location
In this work we show how a generalized hub location problem, that includes as particular cases well-known hub node and hub arc location problems, can be stated as the maximization of a supermodular function. As a consequence, we obtain worst-case bounds on the performance of a greedy heuristic. Moreover, we present two integer programming formulations for the generalized hub location problem. One of these formulations is derived from the properties of supermodular functions, and involves variables with at most two indices. Computational experiments confirm the efficiency of such formulation.
04:45 PM - 05:10 PM
Algorithmic Advances on the Point Location Problem: an Application to Explicit Model Predictive Control (MPC)
Explicit MPC shifts the burden of solving quadratic optimization problems for optimal control strategies from online to offline, reducing them to point location problems. We study few simple yet effective algorithmic techniques to speed-up the offline pre-computation thus expanding the applicability of explicit MPC to higher dimension/longer horizon problems.