Informs Annual Meeting 2017

TD55

INFORMS Houston – 2017

2 - Managing Rail-truck Intermodal Transportation for Hazardous Materials: A Scenario-based Robust Optimization Model Ginger Yi Ke, Memorial University of Newfoundland, St. John’s, NL, Canada. gke@mun.ca, Arifusalam Shaikh Combining multiple transportation modes, intermodal transportation has been widely used in shipping hazardous materials (hazmat). But the corresponding research is still limited, especially when the planning environment contains uncertainties. Herein we propose a bi-objective robust optimization approach to determine the best shipment plan for transporting hazmat in intermodal networks. Both minmax and regret based techniques are applied and compared to find robust solutions. Managerial insights are generated through discussions of the effectiveness and efficiency of our model. 3 - Fair Risk Distribution for the Multi-mode Hazmat Transport Network Design Problem Pirmin Fontaine, Technical University of Munich, Munich, 80333, Germany, pirmin.fontaine@tum.de, Teodor Gabriel Crainic, Michel Gendreau, Stefan Minner To fairly distribute the risk in hazardous material shipment, we propose a population-based risk definition that evaluates the risk in each population center. Moreover, we propose different objective functions for equilibrating the risk and extend the bilevel Hazmat Transport Network Design Problem by considering several transportation modes. We show that both objectives have a positive convex correlation and therefore a significant improvement in risk distribution can be achieved at the cost of just a small increase in total risk. The cities with high risk benefit from the risk redistribution in the beginning. However, strong equilibrations just penalize cities with low risk. 4 - Considering Dangerous Goods in Container Stowage Problems Kiros Kebedow, PhD Student, Hawassa University, Ethiopia, Hawassa University, 05, Hawassa, 05, Ethiopia, kirosmaths@gmail.com, Johan Oppen Container stowage problems are rich optimization problems with both high economic and environmental impact. These problems are typically decomposed into a master bay planning phase, which distributes containers to bay sections of the vessel, and a slot planning phase, which assigns a specific slot within the bay section to each container. In this talk, we extend existing models for slot planning by considering containers with dangerous goods and containers of length 45’. We show that our model can be solved to optimality in reasonable time using standard software. Chair: Roger Chen, Rochester Institute of Technology, 111 Lomb Memorial Drive, 81-2178, Rochester, NY, 14623, United States, rbcgis@rit.edu 1 - Conditional Value at Risk Model for Rail Transportation of Hazardous Material with Demand Due Date Ginger Yi Ke, Associate Professor, Memorial University of Newfoundland, Faculty of Business Administration, Memorial University of Newfoundland, St. John’s, NL, A1B 3X5, Canada, gke@mun.ca, Kan Fang, Tingting Cheng The conditional Value-at-Risk (CVaR) model is employed to generate route choices for the multi-trip rail transportation of hazardous materials, where each demand has an associated due date and each train can run at various train speeds. We investigate the structural properties of the corresponding CVaR model, and provide analytic methods to determine optimal scheduling and routing plans as well as the train speed at each arc, such that the total transportation cost is minimized. 2 - Economic Analysis of Residential Garbage Collection Operations: Case of the City of Kingsville Sachinkumar Prajapati, Texas A&M.University-Kingsville, 1414, W Santa Gertrudis, Apt#702, Kingsville, TX, 78363, United States, prajapatisachin11@yahoo.com, Joon Yeoul Oh, Amir Gharehgozli The quantity of residential garbage increases every year. Along with that, the cost and time of garbage collection increase. The objective of this research is to find an effective yet efficient way to collect residential garbage. To achieve this goal, first, data on the current garbage collection operations is collected and analyzed. Second, Arena simulation is used to improve the operations for the city of Kingsville. Finally, an economic analysis is performed to compare the current and proposed garbage collection operations. The results show that the proposed garbage collection method is 35%-40% less costly compared to the current method. TD54 362A Waste and Hazardous Material Routing Sponsored: Transportation Science & Logistics Sponsored Session

3 - A Meta-heuristic for the Vehicle Routing Problem with Crossdocking

Changsheng Wu, Student, New York Institute of Technology, Manhattan, NY, 10023, United States, cwu25@nyit.edu, Birasnav Muthuraj This study focuses on developing a meta-heuristic for routing vehicles operated by third party logistics service providers to pick up hazardous materials from manufacturing companies and to deliver these materials to customers through crossdocking facility. This study compares the solutions of this meta-heuristic with the solutions derived from exact methods. 4 - Expected Value of Perfect Information of Demand in Propane Distribution Problem Nichalin S. Summerfield, Assistant Professor, University of Massachusetts Lowell, One University Avenue, Lowell, MA, 01854, United States, nichalin_summerfield@uml.edu, Moshe Dror Recent advancement in sensors allows companies to remotely monitor propane levels in customers’ tanks and plan their daily distribution operations. In this work, we study the expected value of perfect information (EVPI) of demand in Propane Distribution Problem (PDP) to decide if it is profitable to install the sensors. Specifically, we compare the solution cost when the levels of propane are known, against the solution cost when customers’ consumption levels are forecasted. In our setting, split deliveries are allowed and any stockout is replenished by a costly special delivery. We use a state-of-the-art heuristic solution for PDPs that minimizes the expected replenishment costs per year. 5 - Organic Waste Recovery: Routing, Siting and Blending Roger Chen, Assistant Professor, Rochester Institute of Technology, 111 Lomb Memorial Drive, 81-2178, Rochester, NY, 14623, United States, rbcgis@rit.edu, William Armington In the US, an estimated 63 million tons of food product is discarded annually, leading many regions to consider diverting food waste from landfills towards other more sustainable technology alternatives. Heterogeneity and inconsistency in the food waste composition, combined with technology constraints require trade offs in optimizing the recycling process, including the routing of organic waste streams. Within this context, this work considers three class of problems: (i) vehicle routing of organic waste streams; (ii) facility location decisions for organic waste recycling technologies; and (iii) waste blending and composition balance. 362B Topics In Finance II: Behavioral Finance and Economics Sponsored: Financial Services Sponsored Session Chair: Xuedong He, The Chinese University of Hong Kong, Hong Kong, xdhe@se.cuhk.edu.hk Co-Chair: Nan Chen, Chinese University of Hong Kong, Shatin N T, Hong Kong, nchen@se.cuhk.edu.hk 1 - A New Preference Model that Allows for Narrow Framing Xuedong He, The Chinese University of Hong Kong, Room 609 William M.W. Mong Engineering Bldg, Shatin, Hong Kong, xdhe@se.cuhk.edu.hk, Jing Guo Barberis and Huang (2009) propose a preference model that allows for narrow framing and apply it to explain individuals’ attitudes toward timeless gambles and high equity premia in the market. To uniquely define the utility process in this preference model and to yield a unique solution when the model is applied to portfolio selection problems, one needs to impose some restrictions on the model parameters, which are too tight for many financial applications. We propose a modification of Barberis and Huang’s model and show that (i) the modified model admits a unique utility process and a unique solution in portfolio selection problems and (ii) it is more tractable than Barberis and Huang’s model. 2 - Naive Behavior in Continuous-time Mean-variance Model This project is concerned with the naive behavior in continuous-time mean- variance model. The objective is obtain the wealth process and portfolio process if we redo naive behavior at each time. By cutting the time interval into many equal length parts and in each part we use the precommitted strategy. By doing so, we can obtain the corresponding wealth process and portfolio process. When the number of cutting parts goes to infinity, we prove that, under some bounded conditions, both of the wealth process and portfolio process will converge. We also find the limits of both processes. TD55 Lin Chen, Columbia University, New York, NY, 10025, United States, lc3110@columbia.edu, Xunyu Zhou

395