Informs Annual Meeting 2017

POSTER SESSION

INFORMS Houston – 2017

45 - Challenges with Data Collection in Methods to Reduce Clinic Wait Times Rachel Moeckel, Graduate Student, University of Michigan, 504 1/2 N.5th Avenue, Ann Arbor, MI, 48104-1110, United States, rachwold@umich.edu, Amy Cohn, Paula Anne Newman-Casey A chief complaint received at an Ann Arbor clinic concerns long wait times during appointments. A team assembled with goals of reducing wait times and introducing patient education in irreducible periods, thereby enabling the clinic to see patients with better efficiency and effectiveness, increase adherence through education, and ultimately improve the patient experience. To do so, we implemented passive RFID technology, the data from which enables exact and automatic calculation of wait times. However, challenges arose with clinic staff, collaboration of engineering and medicine, and conversion of RFID signals to wait times. Here we explore such matters and provide insight into solutions. 46 - Method of Update Forecast Sales Data for DDMRP Respecting Product Similarity Shuto Tsuchiya, Master, The University of Tokyo, 4-6-1, Komaba, Meguro-ku, Tokyo, 1530041, Japan, shuto555@iis.u-tokyo.ac.jp, Demand Driven Material Requirements Planning (DDMRP) is one of production methods that can flexibly respond to a variety of factors such as supply, operation, and demand in supply chains. Although it has established a short-term strategy policy to decide inventory volume for actual demand in DDMRP, it has not been established on how to make a long-term strategy using demand forecast. Thus, it can be said to be a developing method. In this paper, we devise a method of updating demand forecast by an actual demand that can deal with cases of demand change, seasonal variation, new products, and discontinued products in adopting DDMRP. We also show examples of simulations utilizing actual sales performance data. 47 - Optimal Design of Multi-tiered Drug Formularies in Canada: Multiple Criteria Decision Making Vusal Babashov, Doctoral Student, University of Ottawa, 55 Laurier Ave East, DMS.6115, Ottawa, ON, K1N.6N5, Canada, vusal.babashov@uottawa.ca, Sarah Ben Amor, Gilles Reinhardt We focus on the downstream relationships between pharmacy benefit management and consumers. In this phase, we assume that the set of products and the formulary structure (number of tiers and copay schedule) are fixed. We develop drug-to-tier allocation model that uses multi-criteria decision methods to maximize the expected return of the plan, its coverage, and other objective criteria that take into account operational factors (demand, capacity, inventory, and service level), clinical factors (drug effectiveness, interactions, and indications), and economic factors (cost, utility, and preference). 48 - Night Team Scheduling for Pediatric Inpatient Residents Bassel Salka, Center for Healthcare Engineering and Patient Safety, 1075 Beal Avenue, Ann Arbor, MI, 48109, United States, bsalka@umich.edu We present our development of a decision-support tool that uses linear programming to automate the construction of team assignments for residents rotating on the inpatient ward night team each month in a pediatric hospital. 49 - MAP Estimation in Bayesian Variable Selection Models via MINLP Michael Lindon, PhD Student, Duke University, Durham, NC, 27710, United States, michael.lindon@duke.edu A Bayesian approach to variable selection in regression models formulates a prior over the coefficients as a mixture of a continuous distribution and a point mass at zero. Full Bayesian inference, even with moderate numbers of predictors, is computationally prohibitive due to the combinatorial explosion in the number of models. I demonstrate that MAP estimation in many models can be formulated as a mixed integer nonlinear programming problem for a variety of popular priors over the model space. I also provide strategies for developing heuristic algorithms that are able to quickly find good incumbent solutions, which, when used to warm start the MINLP solver, can enhance performance greatly. 50 - A Dynamic Approach to Improve Chemotherapy Pre-mix Policies Hwon Tak, University of Michigan, 500 S.State St, Ann Arbor, MI, 48109, United States, hwontak@umich.edu The key goal of our project is to reduce patient waiting time by mixing chemotherapy drugs before patients arrive in the system or at earlier stages in the process (i.e. pre-mix). In collaboration with the University of Michigan Comprehensive Cancer Center (UMCCC) pharmacy, we have developed an improved template model based on their current pre-mix policy efforts. We present a data driven pre-mix template generator to customize the pre-mix list that the UMCCC currently uses by day of week. We show that full utilization of our dynamic template reduces waste costs, pharmacy workload and patient waiting time. Mondo Sygiyama, Motoki Tajima, Tsuyoshi Nobata, Miyuki Wakasugi, Yudai Honma, Masayuki Soneda

51 - Dynamic Inventory and Price Controls Involving Unknown Demand Tingting Zhou, Rutgers University, Newark, NJ, United States, tingzhou@rutgers.edu, Michael N.Katehakis, Jian Yang We study adaptive policies that handle dynamic inventory and price controls when the random demand for discrete nonperishable items is unknown. Pure inventory control is achieved by targeting newsvendor ordering quantities that correspond to empirical demand distributions learned over time. On the basis of it, we conduct the more complex joint inventory-price control, where demand- affecting prices are chosen. We identify policies that strive to balance between exploration and exploitation and measure their performances by regrets. 52 - Lifted Polymatroid Inequalities for Mean-risk Optimization with Indicator Variables Hyemin Jeon, University of California Berkeley, Sutardja Dai Hall, 450, Berkeley, CA, 94720, United States, hyemin.jeon@gmail.com, Alper Atamturk We investigate a conic quadratic minimization problem with 0-1 indicator variables that arises in mean-risk optimization. Observing that the problem reduces to a submodular function minimization under the binary restriction on the continuous variables, we derive three classes of strong convex valid inequalities by lifting the polymatroid inequalities on the binary variables. We report the result of computational experiments on mean-risk problems with fixed charges and cardinality constraints, which demonstrate the effectiveness of the inequalities in strengthening the convex relaxations and improving the solution times. 53 - Optimal Decentralized Network Formation using a Continuous- time Actor-oriented Model Abhinav Perla, University at Buffalo, Buffalo, NY, 14260, United States, aperla@buffalo.edu, Alexander Nikolaev, Eduardo Pasiliao We address a network formation problem in a decentralized context. The network actors (nodes) are assumed to be independent decision-makers that pursue the same objective - to build a network whose global structure has desirable properties, based on connectivity and closeness. Each actor has partial information about the network structure: the structure of its connections with peers and an estimate of its centrality. We present a continuous-time actor- oriented model as a form of solution and analyze the steady state distribution over the networks formed under this model. Then, we present an approach to infer the model parameters to achieve network outcomes enabling robust global communication. 54 - Research on Logistics Efficiency of the E-commerce with the Influence of the Storage Information Visualization Online Si Chen, Assistant Professor, Southwest Jiaotong University, #111 North of Erhuan Road, ChengDu, 610031, China, chensi@swjtu.edu.cn, Yu Cai, Gang Wu, Mi Gan The effective operation of E-commerce logistics system is the key guarantee for the operation. The efficiency of logistics system can be influenced by many factors. Among all the influences, the improvement of logistics information can be significantly improved with low investment. To study the impact of E-commerce logistics system efficiency from the product storage information visualization. An empirical analysis with DEA model has been used to analyze the influence on the JingDong platform from 2009 to 2016. The results show that relatively low product warehousing information visualization investment can bring preferable logistics operational efficiency improvement. 55 - Analysis on the Coordinated Development Between Logistics Industry and the Other Industries in Sichuan Based on Grey Relativity Analysis Xinyuan Li, Southwest Jiaotong university, ChengDu, China, lixinyuan331@qq.com, Si Chen, Yinying Tang, Mi Gan Logistics industry plays an important role in the development of Sichuan economy. We research into the relationship between the logistics industry and the agriculture,manufactory and business, which are the main economic development departments in Sichuan. And we analysis the gray correlation and the coupling degree of logistics industry and the other industries based on the logistics related data of Sichuan province from 2011 to 2016. Therefore, the development of logistic industry is a strong influence to the other industries, and it can provide adaptive supporting environment forthe other industries. 56 - Does Firm Innovation Improve Organization’s Resilience? A Literature Review Sima Sabahi, PhD Student, North Carolina A&T State University, 119 D.Village Lane, Greensboro, NC, 27409, United States, sima.sabahi@gmail.com This research investigates whether firms that are more innovative are also more resilient to supply chain disruptions, and if so, what capabilities of innovative firms support the firms’ resilience. Our findings indicate that a firm with a more innovative environment would be more resilient to disruptions, because innovation helps them fortify capabilities that are also antecedents of resilience. More specifically, we found that the practices of collaboration, leadership, and information sharing, which are all considered as antecedents of both resilience and innovation, would aid innovative firms to be more resilient.

199