Informs Annual Meeting 2017

SC46

INFORMS Houston – 2017

SC46

a new integrated scheduling problem derived from rolling section of steel production, with consideration of batching decision and demand-side management. The problem is formulated as a mixed integer nonlinear optimization model and then reformulated to a MILP model. Numerical experiments are conducted and results show the model is efficient.

360E Operations Analytics and Optimization for Manufacturing, Logistics and Energy Systems Invited: Project Management and Scheduling Invited Session Chair: Lixin Tang, Northeastern University, Northeastern University, Shenyang, 110004, China, lixintang@mail.neu.edu.cn 1 - Enterprise Wide Optimization for Refinery Planning with Coordination of Energy Consumption and Product Quality Lijie Su, Institute of Industrial & Systems Engineering, Northeastern University, Shenyang, China, sulijie@ise.neu.edu.cn, Lixin Tang, Ignacio E.Grossmann This paper focuses on the Enterprise-wide Optimization(EWO) of planning for refinery plant. The process models of refinery units are integrated in the production planning. The properties of material flow are controlled in order to satisfy the international standard for product oils and environmental demand. The energy consumption with the refining process is taken into account, and coordinated with the properties of final oil products. We present a new mathematical formulation for the EWO of planning to maximize the net profit of refinery. We also analyze the net profit of refinery under different variation conditions, like properties of crude oil, energy consumption and standard of product oil. 2 - A Learning-based Differential Evolution Algorithm for Open-order Coil Allocation Problem in the Steel Industry Fei Zou, Institute of Industrial & Systems Engineering, Northeastern University, Shenyang, China, zfhelen6@163.com, Lixin Tang Open-order coil allocation problem is to allocate open-order coils to unfulfilled customer orders for reducing surplus inventory and improving the material utilization. Since the problem is NP-hard, an improved differential evolution algorithm called DEANS is proposed to achieve the near optimal solutions. In the algorithm, construction heuristics are designed to obtain initial solutions and some greedy strategies are proposed to repair infeasible solutions. Mod mutation operators and learning-based mechanisms are adopted to describe and generate the population. Computational experiments on different scale instances verify the effectiveness of the proposed model and algorithm. 3 - Data Analytics-based for Temperature Prediction in Blast Furnace Top System Xiangman Song, Institute of Industrial & Systems Engineering, Northeastern University, NO. 3-11, Wenhua Road, Heping District, Shenyang, 110004, China, sxmars.cloud@gmail.com, Lixin Tang Iron-making process is the biggest environmental loader with the largest energy consumption. From the perspective of data analytics, the relationship between the temperature distribution of furnace burden in the top of blast furnace and the corresponding charge level shape is studied and the linear regression model is proposed. In addition, a space-time Kalman method is proposed to predict the temperature distribution. Utilizing the relationship model and the predicted temperature, the burden charging operations can be obtained to reduce the energy consumption. Furthermore, the real iron-making data are used to test the performance. Experimental results show that the method is effective. 4 - A Multi-stage Modeling Method for Converter Steel Making Process Chang Liu, Institute of Industrial & Systems Engineering, Northeastern University, Shenyang, China, lc1987328@126.com, Lixin Tang This study concerns with dynamic basic oxygen furnace (BOF) steelmaking process and proposes a data-analytics prediction method. According to the analysis of actual field data, a multi-stage modeling method is proposed for BOF steelmaking process. The key idea of the proposed method is to solve the dynamic prediction in the whole BOF steelmaking process. In order to predict temperature and carbon content more accurately in BOF steelmaking process, IPSO-LSSVM is utilized to establish multi-stage prediction models. In addition, the real industrial data are used for testing. Experimental results show that the data analytics prediction method has a better prediction performance. 5 - Integrated Scheduling of Rolling Section in Steel Production with Demand-side Management Shengnan Zhao, Institute of Industrial & Systems Engineering, Northeastern University, Shenyang, China, shengna1@andrew.cmu.edu, Ignacio E.Grossmann, Lixin Tang Electric grid is now using time-of-use pricing strategy on industry customer, which has great impact on the profit of demand side. Involving energy-intensive process, the steel companies can take advantage of processing flexibility to make better use of electric power, and thus reduce energy cost. In this work, we address

SC47

360F Cyber-Security and Sharing Economy Sponsored: EBusiness Sponsored Session

Chair: Gal Oestreicher-Singer, Tel Aviv University, galos@tau.ac.il Co-Chair: Zhepeng Li, York University, zli@schulich.yorku.ca 1 - Incentivize Ride Matching via Repeated Games Shumin Ma, Chinese University of Hong Kong, Shatin, NT, Hong Kong, smma@se.cuhk.edu.hk, Qi Wu Subsidizing drivers generates a supply multiplier effect at the cost of a firm. It contrasts with surge pricing policies, where excessive demand is accommodated at the cost of riders through price hikes, threatening a firm’s market share. Supply- side incentives thus provide an alternative for a firm to respond to demand surges when increasing fare price is not an option. We take a game approach and model the spatial matching problem as a repeated Stackelberg game. We show that through driver-targeted loyalty programs, the optimal throughput in equilibrium is attained at the optimal amount of incentives. Incentivize excessively, however, would be disruptive as matching ratios will decline. 2 - Comparing Two Solution Seeking Mechanisms in the Online Labor Market Jiahui Mo, Nanyang Technological University, 19251 Preston Road, Apt 2520, Dallas, TX, 75252, United States, jhmo@ntu.edu.sg In the online labor market, firms can seek project solutions via two auction mechanisms: procurement auction and tournament. It is unclear for firms which mechanism to use when they look for better solutions with the same budget. This research compares these two mechanisms and provides some empirical evidences on the effectiveness of these mechanisms. Our findings can provide insights for firms to choose the proper mechanism in the online labor market. 3 - Hotel Strategies to Influence Sales: Incumbents Reaction to Competition from Airbnb Cheng Nie, University of Texas-Dallas, Richardson, TX, United States, cxn123430@utdallas.edu, Eric Zheng, Sumit Sarkar The emergence of Airbnb provides an untapped wealth of supplies to the accommodation market, posing a threat to the incumbent hotels. What should hotels do to improve market performance in face of Airbnb? We consider four strategies that hotels could use to improve their sales: manipulate review, improve quality, reduce price, respond to review. We empirically investigate the effectiveness of these strategies by combining data from Airbnb, Expedia, TripAdvisor, the Texas Comptroller office, and Smith Travel Research.

SC48

361A Operations Systems with Worker Discretion Sponsored: Behavioral Operations Management Sponsored Session

Chair: Hummy Song, The Wharton School, University of Pennsylvania, Philadelphia, PA, 19104, United States, hummy@wharton.upenn.edu 1 - Shared Medical Appointments: An Innovative Approach to Healthcare Delivery Nazl I. Sonmez, PhD Student, London Business School, London, NW1 4SA, United Kingdom, nsonmez@london.edu, Ryan Buell, Kamalini Ramdas In this paper, we examine shared medical appointments (SMAs) as a substitute for one-on-one appointments. Under this innovative approach, a group of patients with similar chronic conditions meet with a doctor simultaneously. We conduct a randomized controlled trial at the Aravind Eye Hospital’s Glaucoma Clinic to assess the effectiveness of shared medical appointments. Preliminary results obtained with the data we have so far suggest that the knowledge and satisfaction level of patients who attend shared medical appointments is significantly higher than that of patients who attend one-on-one appointments.

84