INFORMS 2021 Program Book

INFORMS Anaheim 2021

MD44

2 - Comparing 2-phase and 3-phase Algorithm for Handling Vehicle Routing Problem Jihyun Jo, Pennsylvania State University, University Park, PA, United States, Soundar Kumara In our previous studies, we found out that node density can affect the solution quality of the vehicle routing problem when we use genetic algorithm-based Route-First, Partition-Second method. This method works fine when the node density is high, however, it is not good for the cases with relatively low node density with thousands of visiting nodes in a network. To overcome this problem, we proposed to solve the problem with the nearest neighborhood search algorithm-based Route-First, Partition-Second method and add re-optimization process with genetic algorithm. In this study, we set the different node densities in the same region and investigate the solution qualities of each method. 3 - A Generalization of the Pickup and Delivery Problem with Transshipment and Occasional Drivers Zefeng Lyu, The University of Tennessee at Knoxville, Knoxville, TN, United States, Andrew Junfang Yu A pickup and delivery problem with transshipment, occasional drivers, and predefined origins and destinations is presented. The occasional drivers are getting increasingly common but have yet been fully studied, especially when associated with predefined origin-destination pairs. Although some studies have retained the possibility of distinguishing starting and ending points, they have not evaluated it in experiments. We prove that it is more beneficial of allowing transshipment for the pickup and delivery problem when the drivers have different origins and destinations. A MIP model and a heuristic for large-scale instances are presented, which is a generalization of the existing model. 4 - A Parallel Large Neighborhood Search Algorithm for Vehicle Routing Problems with Asymmetric Costs Keyju Lee, Korea Aerospace University, Goyang-City, Korea, Republic of, Junjae Chae, Younshik Chung This study aims to solve capacitated vehicle routing problem with asymmetric costs. A Large Neighborhood Search (LNS) based algorithm is proposed. A nosing method that efficiently alters the closest customer node from each customer node is developed in the algorithm. The performance of the algorithm outperforms most open source VRP solvers in solving time and quality of solution. 5 - A Decision Making Tool for the Last Mile Delivery Strategy Raghavan Srinivasan, North Dakota State University, Fargo, ND, United States, Joseph Gerard Szmerekovsky, Satpal Singh Wadhwa Last mile delivery has experienced growing interest in utilizing the available capacity from logistics services offered by local people of a region, i.e. crowd logistics. The objective here is to provide a decision-making tool, to determine the optimal balance of deliveries by full time employees, part time employees, and crowd sourcing. A newsvendor type solution is derived for a stylized model incorporating seasonal demand, available capacity, and delivery cost to determine the optimal cost of last mile delivery using full time, part time and crowd sourcing logistics. A heuristic approach based on the stylized model is then provided to determine the optimal last-mile delivery strategy. 6 - Repair Crew Routing on Power Network Restoration Bahar Cavdar, Texas A&M University, College Station, TX, United States, Qie He, Feng Qiu We consider a power distribution network under disruption and study how to efficiently restore the power network where faults occur in some locations and result in outage over the network. Considering the dependencies between the power network and the road network, we develop a bi-directional dynamic programming-based solution method to minimize the total service disruption time for all customers. MD46 CC Room 213D In Person: Recent Advances in Optimization Software I General Session Chair: Hans Mittelmann, Arizona State University, Tempe, AZ, 85287- 1804, United States 1 - Latest Developments in the Artelys Knitro Optimization Solver Richard Waltz, Artelys, Los Angeles, CA, 90045-2603, United States Artelys Knitro is the premier solver for nonlinear optimization problems. Knitro offers both interior-point and active-set algorithms for continuous models, as well as tools for handling problems with integer variables and other discrete structure. This talk will highlight the latest developments in Knitro, focusing on some of the recent advances in solving mixed-integer nonlinear problems. We will also present some Knitro results on large-scale models from various applications. In particular, we will highlight Knitro performance on a selection of large logistic regression models, and on large-scale optimal power flow models from the recent ARPA-E Grid Optimization 2 Competition.

MD44 CC Room 213B In Person: Economic Modeling Contributed Session Chair: Ankita Srivastava, Oklahoma State University, Tulsa, OK, 74137, United States 1 - Optimal Price Subsidy for Plant-based Meat Toward a Differential Game Model Jie Qu, University of Wisconsin Milwaukee, Milwaukee, WI, United States Dealing with environment and food crisis, popularize the fake meat is a possible solution. However, the high production cost made fake meat lack of competitiveness. High cost will be overcome by technology advancement and economics of scale if it, in early market, received subsidy and help from government to become economically competitive. This paper deals with the determination of optimal pricing policy for the firm and optimal subsidy for the government in the monopoly and oligopoly market using differential game. 2 - Digital Borders, Spatial Trade Spillovers, and Development Gabriel Bahr, Oklahoma State University, Stillwater, OK, United States, Bryan Hammer, Andy Luse The purpose of this paper is to expand ICT4D literature by investigating the associations between international trade of technology merchandise and development across countries. Using a spatial autoregression model and data on 45 upper-middle and high income countries from 2009 to 2018, we examine the effects of imports and exports of technology driven trade on two measures of development (GDP and HDI). Additionally, we define spatial borders through a trade partner network and discover spillover effects of trade-development on neighboring countries. 3 - HIT Spillovers and Sustained Cooperation Ankita Srivastava, Oklahoma State University, Tulsa, OK, United States, Chenzhang Bao, Dursun Delen Based on the proposed referral network model we study IT spillover effects from ambulatory facilities to hospitals. Using a panel of 13 years with 2,768 US hospitals matched with approximately 30,000 ambulatory facilities, we find a 1% increase in the average EMR adoption of the regional ambulatory clinics can reduce the inpatient cost of the focal hospital by 0.031% (savings of $51,000) in one year and by 0.059% (savings of $98,000) in four years. Our model is robust to endogeneity issues. We also find support for mechanisms where spillover effects are expected to be stronger. The referral network model and empirical evidence can propagate a culture of sustained cooperation among healthcare providers. 4 - Sequential Price Negotiations for Big-ticket Items: Empirical Discovery and Estimation of Predetermined Strategies Abdullah Gokcinar, PhD Candidate, The University of Texas at Dallas, Richardson, TX, United States, Metin Cakanyildirim, Suleyman Karabuk We empirically analyze negotiations between the seller and buyer over the price of a big-ticket item. In a negotiation, the seller and buyer take turns to accept the other’s offer, make a concession from the previous offer, or exit. Empirical results suggest that a player makes concessions following a predetermined negotiation strategy towards a price, and he/she may accept or exit based on the other player’s offers. Following these, we analytically model negotiations to estimate negotiation strategies along with acceptance and exit probabilities. These estimations can help us in revealing latent negotiation characteristics in different player subpopulations. MD45 CC Room 213C In Person: Vehicle Routing I Contributed Session Chair: Bahar Cavdar, Texas A&M University, College Station, TX, 77843, United States 1 - Ground Vehicle and Unmanned Aerial Vehicle Cooperative Delivery with Mothership Charging Hyung Joo Cha, Korea University, Seoul, Korea, Republic of, DongKyun Kim, Joonyup Eun, Taesu Cheong This research proposes a vehicle routing problem using an electric drone (VRP- ED). The traditional vehicle routing problems with drones, in general, assume that drone batteries are replaced by a ground vehicle driver at a rendezvous point. However, in this talk, we propose a new cooperative vehicle routing model, where the drone acquires its electric power from the mothership on board and consumes its electric power during flight. A mathematical formulation for VRP- ED is presented. Due to the complexity of VRP-ED, a heuristic algorithm is also developed and tested.

86