Informs Annual Meeting 2017

WA52

INFORMS Houston – 2017

2 - Assessment of Container Port Efficiency using Data Envelopment Analysis Sonal Jain, University of Houston, Houston, TX, 77204, United States, jsonal@gmail.com, Qianmei Feng Maritime ports have become a major source of trade, with $900+ billion worth of goods moving in and out of United States by containers. Growing shipping capacity and severe congestion in American ports raises a concern on their state of efficiency. An assessment of relative efficiency of container ports can provide insight into a port’s productivity and performance. This study applies the appropriate Data Envelopment Analysis (DEA) models to study changes in efficiency for 15 U.S. container ports. The relationships between inputs and outputs are examined using the available records since 2000. 3 - Modelling the Service Availability and Resource Allocation for Quay Crane Double Cycling via Closed Queuing Networks Qingcheng Zeng, Dalian Maritime University, School of Transportation Management, Dalian, 116026, China, qzeng@dlmu.edu.cn, Xiaoju Zhang, Ping Wang This paper addresses the resource allocation problem with quay crane double cycling in container terminals. The operation processes is formulated by closed queuing networks. A cost-minimizing optimization model is developed to obtain an optimal resource allocation schedule, and the objective function proves to have convex behaviour with one or two optimal solutions. The asymptotic results offer insight into the types of equipment that should be added to increase efficiency. The imbalances of service ability in each operation have a critical effect on equipment utilization and handling efficiency. 361E Energy Contributed Session Chair: Stefan Feuerriegel, University of Freiburg, Freiburg, Germany, stefan.feuerriegel@is.uni-freiburg.de 1 - Design of Flexibility Portfolio to Facilitate Integration of High Shares of Renewable Energy in Europe Violette Berge, Energy Analyst and Optimization Consultant, Artelys, Montreal, QC, Canada, violette.berge@artelys.com By 2030, the share of fluctuating renewable power generation could attain 30%, following EU targets. This raises the need for additional flexibility to balance demand and supply at all times. Although system flexibility can be provided by various technologies, there is no “one-size-fits-all” solution at a national level and the optimal flexibility portfolio depends on the peculiarities of each country. In this context and using its powerful tool Artelys Crystal Super Grid, Artelys carried out a study for the EC that aimed at evaluating the national flexibility needs and the optimal flexibility portfolio of flexibility solutions at the Member State level to ensure a cost-effective RES integration. 2 - Two Level Distribution Network Planning Problem with Distance Limited Intermediate Facilities Ayse Selin Kocaman, Assistant Professor, Bilkent University, WA52 We introduce a new two-level network planning problem which might have wide applications in the planning of centralized energy and water distribution networks for greenfield development. The network connects demand points to a source point via intermediate facilities in the continuous space. The intermediate facilities reside a primary network that has a tree topology rooted at the source. The demand points are connected to the intermediate facilities with a secondary network in star topology rooted at each intermediate facility. The coverage distance limit on the intermediate facilities can be associated with physical loss constraints in energy or water distribution networks. 3 - A New Approach to the Operation of Micro Grids in Rural Areas Jose-Ignacio Munoz-Hernandez, University of Castilla-La Mancha, Edifico Politecnico - UCLM, Avda Camilo Jose Cela, S/N, Ciudad Real, 13071, Spain, joseignacio.munoz@uclm.es, Luis Serrano-Gomez Distribution networks in rural areas are characterized by high seasonal consumption and alterations in the quality of the supply, which affect both the availability of the service and the electrical components connected to it. Moreover, extinction of the feed-in-tariff for renewable energy is generating uncertainty that is leading towards new solutions for maximizing the operating account. This work analyzes the possibility of establishing a new operating model for isolated micro grids, in rural areas, powered by renewable energy sources. Endustri Muhendisligi, Ankara, 06800, Turkey, selin.kocaman@bilkent.edu.tr, Kagan Gokbayrak

4 - Applying a Game Theory Approach for a Dynamic Energy Bilateral Contract Scheduling Reinaldo Crispiniano Garcia, Associate Professor, University of Brasilia - UnB, SQS.403, BLOCO.R, APT. 106, Brasilia, 70237-180, Brazil, rcgar@yahoo.com, Javier Contreras, Janiele Custodio, Felipe Toledo, Paulo Vinicius Da Cunha, Matheus Barbosa In deregulated electricity markets, contract parties (buyers and sellers) aim to deliver electricity in time intervals during the contract period to obtain the highest possible profits. The parties can operate in spot and retail energy markets. Since this problem includes different parties trying to maximize their own profit, a game theory approach is implemented for the contract settlement. A numerical example shows the properties of this optimization problem for contract scheduling. Moreover, the contract value and the contract volumes are determined to maximize the returns of the involved parties. 5 - Minimization of Investment Costs for Designing Electricity Distribution Networks Stefan Feuerriegel, University of Freiburg, Platz der alten Synagoge, Freiburg, 79098, Germany, stefan.feuerriegel@is.uni- freiburg.de, Alexander Schlüter, Gunther Gust, Dirk G. Neumann We minimize the investment costs for constructing or extending electricity distribution networks, that supply electricity to neighborhoods, while complying with physical constraints. This problem encompasses the optimization of both locations and capacities of conductor lines and is, thus, an extension of the NP- hard capacitated shortest spanning arborescence problem. For this reason, we propose a heuristic that decomposes the problem: it first identifies the network layout and then decides upon capacities, followed by a variable neighborhood search. Its performance is detailed in a real-world case study. 361F Dynamic and Stochastic Routing and Scheduling Sponsored: Transportation Science & Logistics Sponsored Session Chair: Debdatta Sinha Roy, University of Maryland, 7699 Mowatt Lane, 3330 Van Munching Hall, College Park, MD, 20742, United States, debsroy@rhsmith.umd.edu 1 - An Exact Approach for the Vehicle Routing Problem with Time Windows and Flexible Delivery Locations Alexander Döge, Technical University of Munich, Munich, Germany, alexander.doege@tum.de, Markus Frey, Rainer Kolisch, Guy Desaulniers We present an extension of the vehicle routing problem (VRP): the VRP with time windows and flexible delivery locations (VRPTW-FL). In the VRPTW-FL, a customer is no longer automatically assigned to a service location. Instead, one of a set of potential service locations has to be chosen, each of which has a specific capacity. In order to solve the VRPTW-FL, we develop a branch-and-price-and-cut algorithm. In our economic analysis, we assess the benefit of location flexibility on a variety of performance metrics. 2 - Multiple Plan Approach for the Dynamic Team Orienteering Problem Emre Kirac, Assistant Professor, University of Houston-Clear Lake, 2700 Bay Area Blvd., MC41, Houston, TX, 77058, United States, kirac@uhcl.edu, Ashlea Milburn In this study, we introduce a new dynamic routing problem, namely the Dynamic Team Orienteering Problem (DTOP) in which some customers are known a priori while others are dynamic, each associated with a profit. The goal is to maximize the sum of collected profits by visiting a set of customer locations within a time limit. This problem arises in several practical applications such as disaster relief, technician, tourist and school bus routing problems. We adapt a Multiple Plan Approach (MPA) to solve the proposed problem. A total of 1161 new DTOP benchmark instances are introduced and solved. 3 - Simulation-based Learning for Dynamic Time Window Allocation in Attended Home Deliveries WA53

Magdalena A.K. Lang, Research Assistant, RWTH Aachen University, Kackertstraße 7, Aachen, 52072, Germany, magdalena.lang@ada.rwth-aachen.de, Catherine Cleophas, Jan Fabian Ehmke

Allocating time slots to attended home delivery requests affects providers’ delivery costs and future acceptable orders. As the travel time per order depends on the complete routing, revenue management controls face the challenge of undetermined capacity consumption on request arrival. Vehicle routing can approximate travel time but does not anticipate customer choice and future controls. Our simulation-based learning approach approximates the opportunity costs of delivery time slot allocation. It informs dynamic allocation decisions when requests arrive. We present results of a computational study.

460