Informs Annual Meeting 2017

WA66

INFORMS Houston – 2017

WA63

2 - Arc Tour Problems with Complex Objectives Mehmet Basdere, Northwestern University, Technological Institute, 2145 Sheridan Road, Evanston, IL, 60208, United States, MehmetBasdere2016@u.northwestern.edu, Karen Smilowitz, Sanjay Mehrotra In this talk, we present a visit order-based approach and corresponding reformulations for time- and sequence-dependent arc tour problems. Our application comes from marathon course design where the aim is to find a valid simple cycle which visits a predetermined subset of landmark streets without preventing the public access to critical facilities in the network. Time-dependent objective functions are used to measure metrics that change over different stages of the race while sequence-dependent ones are used to measure total turn angle and incline change throughout the course. 3 - Comparison of Several Clustering Algorithms for an Inventory Routing Problem We propose several clustering algorithms as a subroutine of an integrated solution approach for a variant of inventory routing problem. In the clustering phase, we generate several (possibly overlapping) clusters and choose a subset of these generated clusters by solving a set partitioning problem. We determine the cost of serving each cluster by a new tour generation-based algorithm. We compare both the clustering algorithms against each other and the new tour generation-based algorithm against a benchmark algorithm in the literature. The proposed algorithm not only finds results with lower optimality gaps but also finds feasible solutions for difficult instances. 4 - An Application of Inventory Routing Problem in Humanitarian Relief Supplies Distribution Ali Ekici, Ozyegin University, Endustri Muhendisligi Bolumu, Nisantepe Mahallesi Ormak Sokak No:13, Istanbul, Turkey, aliekici@gmail.com, Okan Orsan Ozener, Emre Cankaya Motivated by the practices in humanitarian relief chain, we study an application of inventory routing problem in distribution of humanitarian relief supplies. Different from the classical route cost/length minimization objective in routing problems, our objective is to make a fair distribution to the affected areas. In order to address this distribution problem, we propose a three-phase (clustering, routing and improvement) solution approach. In the proposed solution approach, we take the interaction between routing and allocation decisions into account by first solving the routing problem based on an approximate allocation and then improving the allocation based on the routing decision. Okan Orsan Ozener, Ozyegin University, Nisantepe Mh. Orman Sk. No 13 Cekmekoy, Istanbul, 34794, Turkey, orsan.ozener@ozyegin.edu.tr, Ali Ekici Chair: Seyedbehzad Aghdashi, North Carolina State University, 112 Gwinnett Pl, Cary, NC, 27518, United States, saghdas@ncsu.edu 1 - Robust Optimal Management of Highway Pavement a Utilitarian Approach to User vs. Agency Costs Umit Deniz Tursun, University of Illinois, 205 N Mathews Ave., Urbana, IL, 61801, United States, umitdeniz.tursun@gmail.com Management of highway pavement requires considering a very extensive set of structural performance and environmental impacts simultaneously from both agency and user perspectives. Framework for employing mathematical decision modelling to address analytic approach need for addressing life cycle impacts. A constrained optimization formulation with multi-attribute utility function that reflects the users’ and agencys’ incentives serving as the objective function under feasibility constraints is presented. A case study on Illinois Tollway authority network is presented. 2 - Demand Monotonicity of a Pavement Cost Function Used to Determine Aumann-shapley Values in Highway Cost Allocation Saurav Kumar Dubey, PhD Student, University of Tennessee at Knoxville, 1615 Laurel Avenue, Box 203, Knoxville, TN, 37916, United States, skumardu@utk.edu, Alberto Garcia-Diaz Pavement Thickness and Traffic Lanes are two fundamental requirements of a highway design project. Viewing lanes and 18-kip Equivalent Single-Axle Load (ESAL) as two types of players, a pavement cost function is used to determine the average marginal cost for players also known as the Aumann-Shapley values to allocate cost among vehicle classes. This approach results in large number of players for which an efficient form of computing A-S value is needed. This article will derive a compact form for discrete A-S value whose pavement cost function will be proven to have monotonically increasing cost when demand (ESALs) is increased. WA66 371A Transportation Infrastructure Planning Sponsored: Transportation Science & Logistics Sponsored Session

370D Inventory Management in Vehicle Routing Sponsored: Transportation Science & Logistics Sponsored Session Chair: Ingrid Marcela Monroy Licht, McMaster University, 2685 Kent Avenue, Apt 201, Montreal, QC, H3S 1M8, Canada, marcela.monroy@polymtl.ca 1 - Predictive Analytics and Routing Optimization for an Urban Food Delivery Service Alexander Hess, Research Assistant, WHU. - Otto Beisheim School of Management, Burgplatz 2, Vallendar, 56179, Germany, alexander.hess@whu.edu This talk illustrates the application of predictive analytics in conjunction with vehicle routing for an urban food delivery service. In particular, we show how an improved routing decreases the waiting time for the customers. Further, the ETAs for each delivery can be given with greater accuracy. 2 - Dynamic Inventory Routing Problem with Traffic Congestions Gitae Kim, Hanbat National University, Dept. of Industrial Management Engineering, School of Engineering, Daejeon, 34158, Korea, Republic of, gitaekimemail@gmail.com Vehicle routing problem (VRP) is a complex problem in operations research. When VRP combines with inventory problem, the complexity increases. The combined problem, so called inventory routing problem (IRP), has been studied in recent decades. Like other problems, there are uncertainties in IRP. To cope with uncertain or dynamic situations, dynamic or stochastic IRP has to be considered. However, most studies for stochastic IRP assume that the demand is stochastic. In dynamic VRP, traffic congestion plays an important role for dynamic conditions. Thus, we suggest a stochastic and dynamic IRP model with traffic congestions. We also provide a solution method and numerical results for the problem. 3 - Inventory Routing Problem with Returnable Containers Ingrid Marcela Monroy Licht, Postdoctoral Fellow, McMaster University, Michael G. DeGroote School of Business, Hamilton, ON, ONL8S, Canada, lichti@mcmaster.ca, Kai Huang We present a problem inspired by the gas supply industry where a company provides customers with filled containers in a finite period. Over the days, the customers release the containers; therefore empty containers are accumulated on customer sites depending on the rate consumption. The containers belong to the company; consequently the empty containers should be collected at customers to re-use. This case is modeled as an inventory routing problem with additional considerations. The supplier carries routing and inventory cost, while satisfies the periodic demands of his customers. A branch and price algorithm is proposed to solve the problem. Some computational experiments are analyzed.

WA65

370F Routing Problems Sponsored: TSL, Freight Transportation & Logistics Sponsored Session

Chair: Ali Ekici, Ozyegin University, Endustri Muhendisligi Bolumu, Nisantepe Mahallesi Ormak Sokak No:13, Istanbul, na, Turkey, aliekici@gmail.com 1 - Dynamic Vehicle Routing Problem with Roaming Delivery Locations Gizem Ozbaygin, PhD Candidate, Bilkent University, Dept. of Industrial Engineering, Ankara, Turkey, ozbaygin@bilkent.edu.tr, Martin W. P.Savelsbergh Motivated by the increasing interest in trunk delivery applications, we study the vehicle routing problem with roaming delivery locations (VRPRDL) in a dynamic setting where there may be deviations from original customer itineraries. We propose a rolling-horizon based solution framework that uses the branch-and- price algorithm developed for the VRPRDL to re-optimize the vehicle routes every time an itinerary update is realized and test its efficiency.

465