Print Email Facebook Twitter Facility Location Models in Emergency Medical Service: Robustness and Approximations Title Facility Location Models in Emergency Medical Service: Robustness and Approximations Author Kerkkamp, R.B.O. Contributor Aardal, K.I. (mentor) Van der Mei, R.D. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department Applied mathematics Programme Applied mathematics Date 2014-05-28 Abstract In emergency medical service (EMS) the use of optimisation models and Operations Research techniques is becoming more common. EMS providers incorporate facility location models and simulation software packages into decision support tools, allowing the extensive evaluation of 'what-if' scenarios. We give a literature survey of facility location models applied to EMS and analyse the properties of one EMS model in particular, namely the Maximal Covering Location problem (MCLP). We analyse the sensitivity of the MCLP to changes in the parameters and design approaches to construct insensitive solutions. Furthermore, we prove performance guarantees for two heuristic solution methods for the MCLP: the Greedy Search and the Swap Local Search. All solution methods are numerically evaluated using generated instances and realistic instances based on The Netherlands. Our main research contributions are as follows. First, we apply Robust Optimisation to EMS optimisation models. We derive and analyse a Robust Counterpart formulation for a general linear constraint under the assumption of a certain polytopal uncertainty structure. Second, we present a constructive proof of the tight performance guarantee for the Swap Local Search. The proof explicitly derives the family of worst-case MCLP instances, which have a certain symmetry. Finally, we perform a thorough computational study for the described methods. This research is performed to conclude the Master in Applied Mathematics at the Delft University of Technology in The Netherlands. It is a cooperative project with CWI in Amsterdam, as part of the REPRO research project on ambulance logistics. CWI is the national research institute for mathematics and computer science in The Netherlands. For more information on the REPRO project, see repro.project.cwi.nl. Subject emergency medical serviceoptimisationrobustnesssensitivity analysisapproximations To reference this document use: http://resolver.tudelft.nl/uuid:22cb1ca1-8dbb-42d0-bbc2-c61a33dc479f Embargo date 2015-05-28 Part of collection Student theses Document type master thesis Rights (c) 2014 Kerkkamp, R.B.O. Files PDF MScTUDThesisKerkkamp.pdf 2.06 MB Close viewer /islandora/object/uuid:22cb1ca1-8dbb-42d0-bbc2-c61a33dc479f/datastream/OBJ/view