Informs Annual Meeting 2017

TC44

INFORMS Houston – 2017

TC42

TC44

360A Material Networks Invited: Invited OR and Advanced Manufacturing Invited Session Chair: Sergiy Butenko, Texas A&M University, College Station, TX, 77843-3131, United States, butenko@tamu.edu 1 - Introduction to Network Analysis in the Context of Materials Networks Vladimir Boginski, University of Central Florida, 12800 Pegasus Dr., Orlando, FL, 32541, United States, Vladimir.Boginski@ucf.edu We discuss several concepts from network analysis and their potential interpretations in networks of materials. In particular, dense clusters in networks, such as cliques and clique relaxations will be addressed from the perspective of materials networks. 2 - Materials Networks: Construction and Structural Properties Vladimir Boginski, University of Central Florida, Orlando, FL, United States, Vladimir.Boginski@ucf.edu, Alexander Veremyev In this talk we well discuss how we construct materials networks based on similarity among their density of states functions and overview their basic structural properties. 360B Models and Algorithms for Radiation Treatment Planning Sponsored: Public Sector OR Sponsored Session Chair: Gino J Lim, University of Houston, Houston, TX, 77204, United States, ginolim@uh.edu Co-Chair: Azin Khabazian, University of Houston, 5465 Braesvalley Dr. Apt 566, Houston, TX, 77096, United States, akhabazian@uh.edu 1 - A New Model and Algorithm for Planning Tomotherapy Delivery Wilmer Henao, University of Michigan, 555 E. William Street, Apt 16C, Ann Arbor, MI, 48104, United States, wilmer@umich.edu Current tomotherapy treatments for cancer tumors are developed assuming instantaneous opening and closing of gantry leaf apertures. Hence, conventional treatment planning spurs an excessive number of leaf events, leading to inaccuracies in dose delivery, and machine attrition. We propose a treatment model that explicitly controls the number of leaf events. The result is a treatment plan that represents dose effects more realistically, but which is modeled by a large-scale combinatorial problem. We therefore implement a fast iterative heuristic algorithm that achieves high-quality results. 2 - Incorporating Linear Energy Transfer in the Optimization of Intensity Modulated Proton Therapy Azin Khabazian, University of Houston, Houston, TX, 77096, United States, akhabazian@uh.edu, Wenhua Cao, Gino Lim, Pablo Yepes, David Grosshans, Radhe Mohan We developed a new optimization approach for incorporating linear energy transfer (LET) in intensity modulated proton therapy (IMPT) treatment planning. We hypothesize that the modified objective function will produce IMPT plans that not only satisfy clinical dose criteria but also achieve reduced LET distributions (thus lower biologically effective dose distributions) in critical structures and increased LET in target volumes compared to plans created based on conventional objectives. The resulting plan is compared with the conventional dose optimization regarding dose, LET and the product of dose and LET distributions 3 - Impact of Robust Optimization on Variable Relative Biological Effectiveness in IMPT Dose Distributions Xuemin Bai, University of Houston, Houston, TX, 77054, United States, baid8432@gmail.com, Wenhua Cao, Hans-Peter Wieser, Gino J. Lim, Radhe Mohan Robust optimization has been adopted to ensure robustness of intensity modulated proton therapy (IMPT) treatment plans in face of physical uncertainties such as range and setup errors. However, it has been established using a simplified constant value of relative biological effectiveness (RBE) for protons instead of using more complex and accurate variable RBE models. In this work, we investigate the impact of robust optimization on biologically effective IMPT dose distributions using a recently published RBE model for one phantom case and two head and neck cancer patient cases. TC43

360C Operations Management/Marketing Interface Contributed Session Chair: Refik Gullu, Bogazici University, Istanbul, Turkey, refik.gullu@boun.edu.tr 1 - Optimizing Revenue of Share-private Parking System Shuo-Yan Chou, The National Taiwan University of Science and Technology, Office: MA217, No. 43, Section 4, Keelung Rd, Taipei City, 106, Taiwan, sychou@mail.ntust.edu.tw This study proposes a share-private parking system to increase the availability of parking spaces by optimizing the utilization rate of the existing parking spaces. Through this system, parking spaces owners share their parking spaces when they do not use it and generate revenue. On the other hand, drivers can easily find the available parking space and make a reservation. The main objective of this system is optimizing the revenue from sharing without disturbing the parking owner needs. Before releasing the private parking space to public, the proposed system analyzes the usage pattern of parking owner to find optimal acceptance rate for incoming reservation. 2 - Multiproduct Price Optimization under the Multilevel Nested Logit Model Hai Jiang, Associate Professor, Tsinghua University, Dept of Industrial Engineering, Tsinghua University, Beijing, 100084, China, haijiang@tsinghua.edu.cn, Rui Chen We study the multiproduct price optimization problem under the multilevel nested logit model. When the price sensitivities are identical within each primary nest, that is, within each nest at level 1, we prove that the profit function is concave with respect to the market share variables. We proceed to show that the markup is constant across products within each primary nest, and that the adjusted markup is constant across primary nests at optimality. This allows us to reduce this problem to an equivalent single-variable maximization problem involving a unimodal function. We also investigate the oligopolistic game and characterize the Nash equilibrium. 3 - On the Dependence of Customer Valuation and Service Requirement Chenguang Allen Wu, Northwestern University, 2145 Sheridan Road, Room C229, Evanston, IL, 60208, United States, allenwu@u.northwestern.edu, Achal Bassamboo, Ohad Perry In many services, the value of service may depend on each customer’s individual service requirement. We therefore consider a queueing model in the presence of a stochastic dependence between customers’ valuation and service requirement. We employ a bivariate dependence order to rank the strength of dependence and show that the provider’s revenue is strictly decreasing in the strength of dependence. We demonstrate via numerical examples that the revenue loss caused by a positive dependence can be substantial. Moreover, we show that a positive dependence can also lead to a violation to a known result in the literature that the provider always prefers a large market size when there is no dependence. 4 - Capacitated Assortment and Price Optimization under the Nested Logit Model Rui Chen, Tsinghua University, Shunde Building of Tsinghua University, Beijing, China, chenr15@mails.tsinghua.edu.cn, Hai Jiang We study the capacitated assortment and price optimization problem, where a retailer sells categories of substitutable products subject to a capacity constraint. The goal of the retailer is to determine the subset of products as well as their selling prices so as to maximize the expected revenue. We model the customer purchase behavior using the nested logit model and formulate this problem as a non-linear binary integer program. We then propose an efficient algorithm to obtain its $\epsilon$-approximate solution based on solving a series of multiple- choice parametric knapsack problems. 5 - Modeling Correlation in Competitive Revenue Management Refik Gullu, Professor, Bogazici University, Bogazici University, Industrial Engineering Department, Istanbul, 34342, Turkey, refik.gullu@boun.edu.tr, Engin Yildiz, Taner Bilgic We model a competitive revenue management game where two firms choose their prices to maximize their revenues. Respective valuation of the firms by consumers (consumers’ willingness-to-pay) are correlated random variables. A consumer who is willing to pay a higher than average price to a product or service may be willing to pay a lower or higher than average price for its competitor. We discuss structural properties of the response function, show the existence of unique price equilibrium, and present interesting comparative statics of the game for the case where the willingness-to-pay random variables take a particular bivariate exponential distribution.

351