Informs Annual Meeting 2017

TC62

INFORMS Houston – 2017

TC62

TC63

370C Bayesian and Maximum Entropy Reilability and Operations Models Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Ehsan S. Soofi, University of Wisconsin-Milwaukee, Milwaukee, WI, 53201, United States, esoofi@uwm.edu Co-Chair: Refik Soyer, George Washington University, Washington, DC, 20052, United States, soyer@gwu.edu 1 - Information Theoretic Multistage Sampling Framework for Medical Audits Tahir Ekin, Texas State University, 601 University Dr. Mccoy 451, San Marcos, TX, 78666, United States, t_e18@txstate.edu, Rasim M. Musal The sampling resource allocation decisions for medical audits are crucial and challenging because of the large payment amounts and heterogeneity of the claims. This talk introduces an iterative stratified sampling method that uses Lindley’s entropy measure to evaluate the expected amount of information. We use U.S. Medicare Part B claims outpatient payment data and investigate the versatility of the framework for different overpayment scenarios and resource allocation designs. The proposed method results in reasonable coverage and lower estimation errors while outperforming Neyman Allocation. The framework also can be used to make probability statements on variables of interest. 2 - New Maximum Entropy Newsvendor Models Mahsa Mardikoraem, University of Wisconsin-Milwaukee, 1560 N Prospect Ave, Milwaukee, WI, 53202, United States, mardiko2@uwm.edu, Amirsaman Hamzeh Bajgiran, Ehsan S. Soofi We use the maximum entropy procedures that provide new models for the distribution of demand in the newsvendor problem. Our proposed models include the generalized error family with members such as the normal and Laplace distributions. We also include partial information about the demand provided by under and over stock costs in the maximum entropy calculation via local information constraints. This formulation gives change point demand distributions. Examples include two-piece demand distributions with uniform, exponential, or truncated normal for demand below the order quantity and exponential or truncated normal for demand above the order quantity. 3 - Entropy of Stock-out Distribution and Bayes Risk of the Mean Stock-out Ehsan S.Soofi, University of Wisconsin-Milwaukee, Milwaukee, WI, United States, esoofi@uwm.edu, Majid Asadi, Nader Ebrahimi, Khosro Pichka We study uncertainty and prediction of the stock-out as functions of the inventory level. The distribution of the stock-out, given the inventory, is the residual of the demand. The residual entropy of the stock-out measures the uncertainty and the mean stock-out is the optimal prediction under the quadratic loss. We present applications of the available results for the residual entropy and the mean residual to the stock-out and give new results. Considering variation of the inventory level, we use the expected entropy to measure anticipated uncertainty about the stock-out and compute Bayes risk of the mean stock-out. 4 - Augmented Probability Simulation for Accelerated Life Test Designs Refik Soyer, PhD, George Washington University, Washington, DC, United States, soyer@gwu.edu, Nicholas Polson Designing accelerated life tests presents a number of conceptual and computational challenges. We propose a Bayesian decision-theoretic approach for selecting an optimal design, and develop an augmented probability simulation approach to obtain the optimal design. The notion of a “dual utility probability density” enables us to invoke the concept of a conjugate utility function. For accelerated life tests, this allows us to construct an augmented probability simulation which simultaneously optimizes and calculates the expected utility. To illustrate our methodology, we consider a single-stage accelerated life test design; our approach naturally extends to multiple stage designs.

370D Transportation, Operations Contributed Session Chair: Monia Rekik, Laval University, 2325 rue de la Terrasse, Quebec, QC, G1V 0A6, Canada, monia.rekik@cirrelt.ca 1 - Non Price Reactions to Price Competing Entry Amirhossein Alamdar Yazdi, PhD Candidate, University of Massachusetts, Isenberg School of Management, 121 Presidents Drive, Amherst, MA, 01003, United States, aalamdaryazd@som.umass.edu I study non-price responses to price competition in the airline industry. I look at how incumbent carriers adjust service quality and supply in response to low-cost carrier entry. Our initial results show that incumbent firms service quality improves in the threat-of-entry period but worsens after the entrance of a low- cost carrier. 2 - Sequencing Triple-spreader Crane Operations: Mathematical Formulation and Genetic Algorithm Shabnam Lashkari, PhD. Student in Industrial Engineering, University of Wisconsin-Milwaukee, 1559 N Prospect Avenue, Apt 202, Milwaukee, WI, 53202, United States, lashkari@uwm.edu, Matthew Petering, Yong Wu This paper investigates the problem of scheduling a triple-spreader (i.e. tandem- lift) crane when lifts are subject to a weight limit. We formulate the problem as an integer linear program and develop a fast method for computing a lower bound on the optimal objective value. In addition, we devise a genetic algorithm that produces high quality solutions for small, medium, large, and very large problem instances. 3 - Local Container Drayage Problem under a New Operation Mode Zhaojie Xue, Shenzhen University, Room A423, Civil Engineering Building,, Shenzhen University, Nanshan District, Shenzhen, China, zjxue@szu.edu.cn This study examines the Local Container Drayage Problem (LCDP) under an operation mode in which a tractor can be detached from its companion trailer and assigned to a new task. We have incorporated a set of temporal constraints into the classical VRP to realize this operation by utilizing the idle time available to tractors and coordinating the empty containers moving between customers. A combinatorial benders’ cuts algorithm and a tabu search algorithm are proposed. Some numerical experiments are conducted to assess the performance of the proposed algorithm, quantify the benefit of the new operation mode, and identify the conditions under which the mode is effective. 4 - A Robust Approach to Airport Gate Assignment with a Solution-dependent Uncertainty Budget Chao Zhang, Sun Yat-Sen University, No. 135, Xingang Xi Road,, Zhenghuantang B509, Guangzhou, China, chaozhang1209@gmail.com, Liang Xu, Fan Wang, Feng Xiao Airport gate assignment (AGA) aims to assign flights to gates according to their arrival and departure times. To tackle flight delays in airports, we propose a robust airport gate assignment (RAGA) to maximize the probability that the total real- time gaps between consecutive flights at the same gate is no less than a pre-specified target. Then we develop an asymptotically tight upper bound for the violation probability. Based on the upper bound, a solution-dependent uncertainty budget is introduced to develop a robust counterpart for the RAGA. Empirical study on ShuangLiu International Airport (CTU) indicates that our proposed robust approach for AGA outperforms existing methods. 5 - Combinatorial Bid Generation for Tl Services Procurement Monia Rekik, Associate Professor, Universite Laval, 2325 rue de la Terrasse, Quebec, QC, G1V. 0A6, Canada, monia.rekik@cirrelt.ca, Farouk Hammami, Leandro C. Coelho We address the Bid Generation Problem in combinatorial auctions for the procurement of TL transportation services. The objective is to identify profitable transportation contracts to bid on based on the carrier existing network and other operational constraints. An Adapted Large Neighborhood Search heuristic is proposed. Computational results show that the ALNS heuristic performs well in terms of CPU time and solution quality.

358