The Experts below are selected from a list of 374724 Experts worldwide ranked by ideXlab platform
Can Akkan - One of the best experts on this subject based on the ideXlab platform.
-
search algorithms for improving the pareto front in a timetabling problem with a Solution Network based robustness measure
Annals of Operations Research, 2019Co-Authors: Gulcin Ermis, Can AkkanAbstract:We develop search algorithms based on local search, and a matheuristic that solves a set of mixed integer programming models to improve the robustness of a set of Solutions for an academic timetabling problem. The matheuristic uses the Solution pool feature of CPLEX while solving two related MIP models iteratively. The Solutions form a Network (Akkan et al. in Eur J Oper Res 249(2):560–576, 2016. doi: 10.1016/j.ejor.2015.08.047), in which edges are defined by the Hamming distance between pairs of Solutions. This Network is used to calculate a robustness measure, where disruption of a Solution is assumed to occur when the time slot to which a team had been assigned is no longer feasible for that team and the heuristic response to this disruption is choosing one of the neighbors of the disrupted Solution. Considering the objective function of the timetabling problem and this robustness measure results in a bi-criteria optimization problem where the goal is to improve the Pareto front by enlarging the Network. We compare the performance of the heuristics on a set of random instances and seven semesters’ actual data. These results show that some of the proposed local search algorithms and the matheuristic find high quality approximate Pareto fronts. Besides being one of the few timetabling algorithms in the literature addressing robustness, a key contribution of this research is the demonstration of the effectiveness of the matheuristic approach. By using this matheuristic approach, for any discrete optimization model that can be solved optimally or near-optimally in an acceptable time, researchers can develop a robustness improvement algorithm.
Philip N Klein - One of the best experts on this subject based on the ideXlab platform.
-
the two edge connectivity survivable Network problem in planar graphs
International Colloquium on Automata Languages and Programming, 2008Co-Authors: Glencora Borradaile, Philip N KleinAbstract:Consider the following problem: given a graph with edge-weightsand a subset Qof vertices, find a minimum-weight subgraphin which there are two edge-disjoint paths connecting every pair ofvertices in Q. The problem is a failure-resilient analogof the Steiner tree problem, and arises in telecommunicationsapplications. A more general formulation, also employed intelecommunications optimization, assigns a number (orrequirement) rve{0,1,2} to each vertex vin the graph; for each pairu,vof vertices, the Solution Network is requiredto contain min{ru,rv} edge-disjointu-to-vpaths. We address the problem in planar graphs, considering a popularrelaxation in which the Solution is allowed to use multiple copiesof the input-graph edges (paying separately for each copy). Theproblem is SNP-hard in general graphs and NP-hard in planar graphs.We give the first polynomial-time approximation scheme in planargraphs. The running time is O(nlogn). Under the additional restriction that the requirements are in{0,2} for vertices on the boundary of a single face of a planargraph, we give a linear-time algorithm to find the optimalSolution.
Hussein Naseraldin - One of the best experts on this subject based on the ideXlab platform.
-
facility location a robust optimization approach
Production and Operations Management, 2011Co-Authors: Opher Baron, Joseph Milner, Hussein NaseraldinAbstract:In this research, we apply robust optimization (RO) to the problem of locating facilities in a Network facing uncertain demand over multiple periods. We consider a multi-period fixed-charge Network location problem for which we find (1) the number of facilities, their location and capacities, (2) the production in each period, and (3) allocation of demand to facilities. Using the RO approach we formulate the problem to include alternate levels of uncertainty over the periods. We consider two models of demand uncertainty: demand within a bounded and symmetric multi-dimensional box, and demand within a multi-dimensional ellipsoid. We evaluate the potential benefits of applying the RO approach in our setting using an extensive numerical study. We show that the alternate models of uncertainty lead to very different Solution Network topologies, with the model with box uncertainty set opening fewer, larger facilities. Through sample path testing, we show that both the box and ellipsoidal uncertainty cases can provide small but significant improvements over the Solution to the problem when demand is deterministic and set at its nominal value. For changes in several environmental parameters, we explore the effects on the Solution performance.
Gulcin Ermis - One of the best experts on this subject based on the ideXlab platform.
-
search algorithms for improving the pareto front in a timetabling problem with a Solution Network based robustness measure
Annals of Operations Research, 2019Co-Authors: Gulcin Ermis, Can AkkanAbstract:We develop search algorithms based on local search, and a matheuristic that solves a set of mixed integer programming models to improve the robustness of a set of Solutions for an academic timetabling problem. The matheuristic uses the Solution pool feature of CPLEX while solving two related MIP models iteratively. The Solutions form a Network (Akkan et al. in Eur J Oper Res 249(2):560–576, 2016. doi: 10.1016/j.ejor.2015.08.047), in which edges are defined by the Hamming distance between pairs of Solutions. This Network is used to calculate a robustness measure, where disruption of a Solution is assumed to occur when the time slot to which a team had been assigned is no longer feasible for that team and the heuristic response to this disruption is choosing one of the neighbors of the disrupted Solution. Considering the objective function of the timetabling problem and this robustness measure results in a bi-criteria optimization problem where the goal is to improve the Pareto front by enlarging the Network. We compare the performance of the heuristics on a set of random instances and seven semesters’ actual data. These results show that some of the proposed local search algorithms and the matheuristic find high quality approximate Pareto fronts. Besides being one of the few timetabling algorithms in the literature addressing robustness, a key contribution of this research is the demonstration of the effectiveness of the matheuristic approach. By using this matheuristic approach, for any discrete optimization model that can be solved optimally or near-optimally in an acceptable time, researchers can develop a robustness improvement algorithm.
Glencora Borradaile - One of the best experts on this subject based on the ideXlab platform.
-
the two edge connectivity survivable Network problem in planar graphs
International Colloquium on Automata Languages and Programming, 2008Co-Authors: Glencora Borradaile, Philip N KleinAbstract:Consider the following problem: given a graph with edge-weightsand a subset Qof vertices, find a minimum-weight subgraphin which there are two edge-disjoint paths connecting every pair ofvertices in Q. The problem is a failure-resilient analogof the Steiner tree problem, and arises in telecommunicationsapplications. A more general formulation, also employed intelecommunications optimization, assigns a number (orrequirement) rve{0,1,2} to each vertex vin the graph; for each pairu,vof vertices, the Solution Network is requiredto contain min{ru,rv} edge-disjointu-to-vpaths. We address the problem in planar graphs, considering a popularrelaxation in which the Solution is allowed to use multiple copiesof the input-graph edges (paying separately for each copy). Theproblem is SNP-hard in general graphs and NP-hard in planar graphs.We give the first polynomial-time approximation scheme in planargraphs. The running time is O(nlogn). Under the additional restriction that the requirements are in{0,2} for vertices on the boundary of a single face of a planargraph, we give a linear-time algorithm to find the optimalSolution.