Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SB76

2 - Application of a Novel PROMETHEE-based Method for Construction of a Group Compromise Ranking to Prioritization of Green Suppliers in Food Supply Chain Milosz Kadzinski, Poznan University of Technology, Piotrowo 2, Poznan, 60-965, Poland We propose a hybrid outranking-based approach for group decision aiding. It combines novel procedures based on Binary Linear Programming for constructing a compromise ranking with robustness analysis accounting for the decision makers’ incomplete preferences. The framework’s applicability is endorsed with a study concerning a food processing industry situated in India. The case company has been supported in selecting the most preferred supplier, while considering economic and green criteria and procuring the environmental friendly policy. The study’s significance derives from a prodigious focus of the food sector, which is constantly gaining in importance in today’s global economy. 3 - Bilevel Programming for Generating Discrete Representations in Multiobjective Optimization A bilevel programming formulation is proposed to identify a Decision Maker(DM)’s most preferred solution without generating the entire nondominated set. The bilevel program is capable of delivering an efficient solution that maps into a given set, provided that one exits. If the DM’s preferences are known a priori, they can be used to specify the given set. Alternatively, we propose a search method using the bilevel program to obtain a representation of the nondominated set when DM’s preferences are not available. The method searches the outcome space using a partitioning scheme. We show that the method finds a representation with a specified coverage error level in a finite number of iterations. n SB75 West Bldg 212B Joint Session MAS/Practice Curated: Cybersecurity Analytics and Applications II Joint Session Chair: Natalie M. Scala, Towson University, Dept. of e-Business and Technology Management 1 - A Risk Model for Voting Systems at Local Precincts Natalie M. Scala, Towson University, Dept. of e-Business and Technology Management, 8000 York Road, Towson, MD, 21252, United States, Paul L. Goethals Cyber, physical, and insider vulnerabilities are threats to the integrity of elections. Most research has predominantly examined cyber threats at the aggregate state level in U.S. elections. Little attention has been paid to insider or human threats, particularly at polling places. This research examines vulnerabilities at polling precincts and develops a risk model, based on Markov chains, to assess total threat. We then identify best practices to mitigate threat and illustrate the application of the model through a case study of a mid-Atlantic state. 2 - Cybersecurity Applications of Blockchain Technology Many are lured to blockchain by the promise of security and privacy in the transaction. However, blockchain security and privacy are mostly a myth. But many applications will still benefit from blockchain technology when reliability and trust are at stake. Cybersecurity is a realm rich with applications as an industry founded on reliability and trust for protecting infrastructure, physical and virtual. This talk will look at the cybersecurity industry and how it can use blockchain to provide better service to clients, public and private. 3 - Defending Against Chained Cyber-attacks by Adversarial Agents: An Optimization Approach Vivin Paliath, Arizona State University, Tempe, AZ, 85281, United States, Paulo Shakarian Cyber-adversaries employ a variety of malware and exploits to attack computer systems. Existing paradigms that model such cyber-adversarial behavior do not account for sequential or “chaining of attacks. We take the first steps toward addressing this need thru a framework that allows for the modeling of sequential cyber-attacks, taking into account complex-interdependencies between vulnerabilities and exploits. The framework identifies the overall set of capabilities gained by an attacker through the convergence of a fixed-point operator. We study the problem of determining an optimal defense strategy to block the attacker from gaining certain capabilities. Gokhan Kirlik, University of Maryland Medical System, 920 Elkridge Landing Rd, Linthicum Heights, MD, 21090, United States, Serpil Sayin James Howard, Lead Data Scientist and Service Manager, The Johns Hopkins University, 11100 Johns Hopkins Road, Laurel, MD, 20723, United States

n SB76 West Bldg 212C Modelling and Control of Production Systems General Session Chair: Baris Tan, Koc University, Istanbul, 34450, Turkey 1 - Integrated Quality and Production Logistics Analysis of Manufacturing Systems with Lead-time Dependent Product Rework Loops Marcello Colledani, Poltecnico di Milano, Milano, Italy, Alessio Angius Products whose quality deteriorates with the time spent within specific portions of a manufacturing system are commonly found in the food, semiconductor, and polymer processing industries. Usually, in these systems, perishable parts must be scrapped if their system sojourn time exceeds a certain threshold. However, there are cases, for example in semiconductor wafer front-end production lines, in which these products can be re-inserted in the system and re-processed to meet the lead-time requirements. This paper presents an analytical method for the exact analysis of the lead-time distribution in these systems, providing the capability to support their design and operation. 2 - Dynamic Pricing and Lead Time Quotation in a Make-to-Stock Queue Baris Balcioglu, Sabanci University, Faculty of Engineering and Applied Sciences, Orhanli Tuzla, Istanbul, 34956, Turkey We study an environment where customers pay more if their demand is instantaneously satisfied. With dynamic pricing and lead-time quotation policies, the manufacturer aims at preventing loss of demand from those who have to wait. As an extension, customers can form two classes: those who are willing to pay more for instantaneous delivery and those who can opt to wait in return for being charged lower prices. We extend the application of the multilevel inventory rationing policy for the manufacturer to serve these different types of customer classes. Via numerical examples, we assess if elaborate dynamic pricing and lead- time quotation policies increase profit when compared to simpler policies. 3 - Data-Driven Control of a Production/Inventory System by Using Marking-Dependent Threshold Policy Baris Tan, Koc University, Rumeli Feneri Yolu, Sariyer, Istanbul, 34450, Turkey, Siamak Khayyati We introduce the Marked Markov Arrival Process (MMAP) framework to model a sproduction ystem that generates different signals (markings) based on the system status. We propose a marking-dependent threshold policy to control the system with partial information. We propose a matrix analytical approach to analyze a production/inventory system with MMAP information and demand arrivals and production times that follow a Markov Arrival Process. We then use a joint simulation and optimization approach to determine the parameters of the threshold policy by using the shop-floor data, the processing time data of the machine and its statistical information, and the inventory position. 4 - Continuous Simulation Optimization with Model Mismatch Giulia Pedrielli, Arizona State University, AZ, United States, Alireza Inanlouganji, Georgios Fainekos, Sebastian Pokutta Multi-fidelity simulation optimization uses low-cost simulation models to optimize functions that can be only evaluated through an expensive, black-box oracle. Current methods are either applicable to discrete problems or suffer from the curse of dimensionality. We propose a novel multi-fidelity continuous global optimization algorithm that works with general low fidelity models, and we use the state of the art Efficient Global Optimization algorithm as a benchmark. Results show that the proposed algorithm can achieve two orders improved accuracy in optimal location identification. 5 - Sales and Operations Planning for Product Rollovers Under Limited Capacity A product rollover takes place if a product generation is replaced by its successor. The key decisions are the offered amount of each product generation and their pricing. If the production of the old generation is stopped prior to its withdrawal from the market, an inventory decision has to be made additionally. In contrast to the existing literature, we consider finite production capacities. We propose a stylized model that captures the firm’s decisions and the substitution behavior of the customers. Structural insights with respect to the impact of limited production capacity based on analytical solutions and a numerical study are discussed. Extensions to a stochastic setting are outlined. Justus Arne Schwarz, University of Mannheim, Schloss Schneckenhof Ost, Mannheim, 68131, Germany, Baris Tan

59