Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

POSTER SESSION

62 - Analysis on the Competition Between Inland Ports Under Carbon Emission Constraints Qinglan Huang, Southwest Jiaotong University, Chengdu, 610031, China, Mi Gan, Ronghui Xie With the progress of One Belt And One Road strategy, the construction of inland ports has become more and more popular, and the competition between them has become increasingly fierce. But reducing carbon dioxide emissions is one of the most important tasks of the 21st-century society. Therefore, this paper establishes a Hotlin model to describe the competitive relationship between the two inland ports, then respectively analyze the competition when one port price increasing or both inland ports price increasing under the limitation of the cap and cap-and- trade carbon emissions policy. For enterprises and the government, it also provides a basis for decision-making. 63 - Data Driven Equilibrium Pricing and Ordering in a Multi Leaders One Follower Game Yong Ge Yang, National Tsing Hua University, Phoenix Arizona, Hsin Chu, 30071, Taiwan This study constructs a multi-leader-one-follower game-theoretic model of supply chain. We consider three roles in a supply chain, which are one supplier, multiple retailers, and aggregate consumer. Retailers will be the leaders. Consumer play as the follower. Leaders keep conjecturing about the demand of followers. Retailers face the uncertain demand. We investigate the Nash equilibrium of the price competition between retailers with pricing and ordering decisions. We embed this game to a data-driven algorithm with censored demand. 64 - Vertical Information Acquisition and Horizontal Information Sharing Strategy in a Supply Chain Jianghua Wu, Renmin University of China, School of Business, Beijing, 100872, China, Fan Jiang We study the vertical information acquisition and horizontal information sharing strategy in a two-echelon supply chain, which consists of one upstream manufacturer and two downstream retailers. The study shows that retailers have no incentives to share information with manufacturer voluntarily, and we identify conditions under which full information acquisition can be achieved as long as with information acquisition the total supply chain is better off. Moreover, we show that market variation, boundary equilibrium and the competition mode (Cournot or Bertrand) in the downstream play a significant role in determining whether the competing retailers will share information with each other. 65 - Delivering Big and Bulky Items Quickly Melissa Quinton, University of Arkansas, Fayetteville, AR, United States, Olivier Kwizera, Yue Wang, Joseph Geunes, Sarah G. Nurre One factor influencing the shift from retail store to online shopping is the availability of quick and cheap shipping. Notably, Amazon prime members receive small to medium sized items in 2 days for free. We examine the necessary supply chain design for quickly delivering big and bulky items, such as appliances and furniture. For different big and bulky segments and delivery services, we determine where and how much inventory should be held to enable guaranteed committed delivery times. We present the trade-off analysis between cost and delivery time by geographic region and demand level. 66 - Conjoint Analysis on the Use of Crowd Sourcing Model in Last Mile Delivery for CEP Services Mingwei Guo, North Dakota State University, UGPTI, Fargo, ND, 58105, United States, mingweiguo@live.com Last mile delivery Crowd-sourcing model would effectively improve the efficiency and mitigate the cost for CEP service providers, but actual market response may differ from our numerical results. From a customer perspective, the quality of the delivery service contains not only time and accuracy, but also the overall experience and level of convenience. If hybrid model is to be used by those customers, the quality and satisfaction of the service may be influenced by the change of staff. This study is designed to evaluate and estimate the direct and indirect impact on customer experience if hybrid models/scenarios are utilized by CEP service providers. 67 - The Conflict Free UGV Routing in an Autonomous Warehouse Mona Issabakhsh, University of Miami, South Miami, FL, 33143, United States, Seokgi Lee High labor manufacturing costs have brought a significant amount of attention to smart robotic warehouse management systems, where tasks are done by autonomous unmanned ground vehicles (UGVs). We present a mathematical UGV routing model with battery recharge schedules, which prevents multiple UGV collisions while minimizing costs of routing and energy consumption. 68 - How Calculative are Managers when Evaluating Signals: An Empirical Examination of Signaling Theory in Trust Formation in a Supply Chain Pook Carson, Associate Professor, Salt Lake Community College, Salt Lake City, UT, 84123, United States Formal models in the management sciences assume a calculative worldview. We experimentally examine how calculative managers are when trusting a supply chain partner. We distinguish between fully-calculative, non-calculative, and signaling/credibility models. The results show that interpretations are influenced by credibility which overwhelms the effect of history. Partner replaceability decreases trust from credible signals as attributions are made to external market

incentives not type. Participants do not distinguish between costly past behavior and bonds as signals. Managers are calculative, but less forward-looking than strong-form calculative approaches suggest. 69 - Using Board Game in Supply Chain Management Classes Ziping Wang, Morgan State University, Baltimore, MD, United States This poster discusses the invention of an instructional board game in supply chain management (patent pending). Data from controlled experiments are studied to show the performance of the board game in supply chain management learning. 70 - Component Procurement and End-product Assembly with Demand and Supply Uncertainty, Using Spreadsheet Based Optimization Models. Ramesh Bollapragada, Professor, San Francisco State University, San Francisco, CA, 94132, United States This paper presents the planning and coordination of component procurement and end-product assembly in a supply chain. Using linear programming models in the framework of Excel spreadsheets, we identify the effect of problem parameters on performance measures such as: inventory, shortage costs, trade- offs involving the total expected costs, the quantity of each component, and end-product to produce in a time horizon. We show through our experimentation that the highest cost drivers are product shortage costs and shared component ordering costs. Further, we show that the component inventory and purchasing decisions are sensitive to lead time increases. 71 - The Spatial Structure Relative Research Between the Development of Regional Industries and the Logistics Demand in Sichuan Province of China Xiaozhen Deng, Southwest Jiaotong University, Chengdu, China, Si Chen, Mi Gan With the evolution of regional industrial structure, the regional logistics demands generate new industrial correlation characteristics of spatial distribution, which are influenced by the spatial and temporal distribution characteristics of the industries. And those characteristics have important reference value for regional industrial layout and allocation of logistics facilities. With 21 cities as the study area and the three industries and the logistics demands of each cities data as data sources. The methods of spatial auto-correlation and regression analysis are used to do spatial and temporal characteristics correlation analysis between logistics demand and regional Industrial. Exhibit Hall Poster Competititon Monday Poster Competition Competition Poster Session Chair: Neng Fan, University of Arizona, 1127 E. James E. Rogers Way, P. O. Box 210020, Tucson, AZ, 85721, United States Co-Chair: Junming Yin, University of Arizona, Management Information Systems Department, McClelland Hall, Room 430BB, Tucson, AZ, 85721, United States Co-Chair: Burcu B. Keskin, University of Alabama, 4138 Meretta Lane, Tuscaloosa, AL, 35406, United States 1 - IoT and Connected Vehicle Analytics for Intelligent Air Quality Management Oleg Gusikhin, Ford Motor Company, Research & Innovation Center,Dearborn, MI, 48121, United States, Omar Makke Air quality is an increasing concern as urban areas continue to grow. For automotive industry, protecting vehicle occupants from pollution is a critical area of competitive advantage. This work illustrates how advancements in IoT sensors, connected vehicles technology and analytics enable intelligent cabin air quality management and optimized filter maintenance. 2 - Directed Exploration in PAC Model-free Reinforcement Learning Min-hwan Oh, Columbia University, New York, NY, 10027, United States, Garud N. Iyengar One of the fundamental challenges in Reinforcement learning (RL) is how to balance exploration and exploitation. We study an exploration method for model- free RL that generalizes the counter-based exploration bonus methods and takes into account long term exploratory value of actions rather than a single step look- ahead. We propose a model-free RL method that modifies Delayed Q-learning and utilizes the long-term exploration bonus with provable efficiency. We show that our proposed method finds a near-optimal policy in polynomial time (PAC- MDP), and also provide experimental evidence that our proposed algorithm is an efficient exploration method.

182