Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TD02

2 - Sensor Network Localization Under Imperfect Information Liang Xu, University of Pittsburgh, 3700 O’Hara Street, 1048 Benedum Hall, Pittsburgh, PA, 15261, United States, Bo Zeng The use of wireless sensors has drastically increased, and accurately estimating the location of sensors under imperfect information becomes more important. In this study, we consider sensor network localization problem through game theory and bilevel generalized interdiction framework. Bilevel optimization models with different distance measures are proposed, and numerical results are also demonstrated. 3 - Security Games with Multiple Adversaries Wei Wang, University of Pittsburgh, 1025 Benedum Hall, Pittsburgh, PA, 15261, United States, Bo Zeng We build a defender-attacker-defender(DAD) model with multiple adversaries, in which different cooperative levels of the attackers are considered. We demonstrate our model on RTS-96 24-bus systems with one and two areas. Computational experiments show that different attackers’ collaboration levels have very much different influences on these systems. The results can help allocate defense resources against one group or cooperative adversaries. 4 - A Study on Approximation Methods of Robust Optimization with Integer Recourse Bo Zeng, University of Pittsburgh, 3700 O’Hara Street, Benedum Hall, Pittsburgh, PA, 15260, United States A few approximation strategies are investigated. Numerical results will be presented to evaluate their qualities. 5 - Traveling Salesman Problem with Probabilistic Visiting Chances Between Nodes Yiwen Xu, Assistant Professor, North Dakota State University, 1410 14th Avenue North, Room 202 Civil & Industrial Engineering, Fargo, ND, 58102, United States, saeid rasti In this talk, we will propose a probabilistic visiting TSP including a success visiting probability from node i to j. A chance constraint is added to guarantee that the success probability of the whole travel has a lower bound. Note that this TSP is different from (1) the PTSP proposed by Jaillet in which only a subset of potential nodes needs to be visited, and (2) the VRP with stochastic demand by Bertsimas. All nodes must be visited exactly once. The problem is formulated as an MILP. A PTAS is proposed by a DP algorithm. n TD02 North Bldg 121B Stochastic Optimization Sponsored: Optimization/Optimization Under Uncertainty Sponsored Session Chair: Ruediger Schultz, University of Duisburg-Essen,Faculty of Mathemati, Thea-Leymann-Str. 9, Essen, D-45127, Germany 1 - Newsvendor in Sobolev Space - Stochastic Shape Optimization Ruediger Schultz, University of Duisburg Essen, Department of Mathematics, Thea-Leymann-Str. 9, Essen, D-45127, Germany The talk reports on some developments in handling shape optimization under uncertainty by means. of recourse stochastic programming. 2 - On Stability of a Risk Averse Linear Bilevel Problem under Stochastic Uncertainty Johanna Burtscheidt, University of Duisburg-Essen, Germany Stochastic bilevel problems arise from the interplay between two decision makers on different levels of hierarchy where the lower level problem is entered by a random vector. We present a risk averse formulation for linear bilevel problems under stochastic uncertainty which are based on special risk measures. In particular, structural properties and qualitative stability of the optimal value function of this model under perturbation of the underlying Borel probability measure will be investigated wrt weak convergence of probability measures. Focusing on a finite number of realizations of an underlying random vector, equivalences to single-level problems conclude the talk.

n TD03 North Bldg 121C Sara Rezaee Vessal Session Sponsored: Technology, Innovation Management & Entrepreneurship Sponsored Session Chair: Antoine Feylessoufi, University of Cambridge, Cambridge, CB2 3BU, United Kingdom 1 - The Carrot or the Stick: Quality in Engineering Contracts Pascale Crama, Singapore Management University, 50 Stamford Road, Singapore, 178899, Singapore, Zhenzhen Chen, Wanshan Zhu In large engineering procurement contracts, the main contractor is responsible for the quality of the end product to the consumer. The main contractor uses subcontractors and the final quality is influenced by the efforts of all the contracting parties. Outcomes that do not meet the minimum required standard require rework, the cost of which is shared between the main contractor and the subcontractor. We find that the first-best may not be attained even for a risk- neutral subcontractor because of the shared rework cost; yet for a risk-averse subcontractor, the main contractor may choose to increase its share of the rework cost. 2 - Personal Fabrication as an Operational Strategy: Value of Delegating Production to Customer Nagarajan Sethuraman, Kenan Flagler Business School, UNC Chapel Hill, Campus Box 3490, Chapel Hill, NC, 27599, United States, Ali Kemal Parlakturk, Jayashankar M. Swaminathan In this paper, we study an operational strategy enabled by 3D printing—- Personal Fabrication (PF) —-in which a firm focuses on product’s design and delegates its production to customers. We characterize the conditions under which, such a strategy benefits the firm. We study the implications of various roadblocks for such a strategy: high production costs of 3D printing, intellectual property concerns and product liability issues. 3 - Technology Adoption in Organisations: An Evolutionary Model Antoine Feylessoufi, University of Cambridge, St Andrews Street, Cambridge, CB2 3BU, United Kingdom Through social interactions, the behaviour of an individual is affected by the population but also influences the other members within that population. In a new approach to capture this effect on technology and innovative practices adoption in organisations, we incorporate social comparison into an evolutionary model widely used in biological ecosystems. We find that unexpected and extreme levels of innovative behaviour (or lack thereof) can emerge in organisations through this mechanism. We also find that for a same technology and firm reward mechanism, the culture of the organisation can lead to different adoption patterns. n TD04 North Bldg 122A Approximation Algorithms Sponsored: Optimization/Integer and Discrete Optimization Sponsored Session Chair: Xiangkun Shen, University of Michigan, Ann Arbor, MI, 48109, United States 1 - Submodular Optimization with Contention Resolution Extensions Benjamin Moseley, Carnegie Mellon University, 5000 Forbes Av, Pittsburgh, MO, 15208, United States This talk considers optimizing a submodular function subject to constraints. Previous work usually (1) discovers a fractional solution to the multi-linear extension and (2) rounds this solution to an integral solution via a contention resolution scheme. Diverging from previous work, we introduce a method called contention resolution extensions. A contention resolution extension combines the contention resolution scheme into a continuous extension of a discrete submodular function. In the case where there is loss in both (1) and (2), by optimizing them together, the losses can be combined resulting in an overall improvement. 2 - Random Sampling and Contraction for Hedge and Hyperedge Connectivity Debmalya Panigrahi, Duke University, Durham, NC, United States In this talk, I will introduce the problem of hedge connectivity which generalizes the notion of (hyper)-edge connectivity in graphs and hypergraphs to model robustness of a network under dependent edge failures. I will then give new algorithms for finding the hedge and hyperedge connectivity of graphs and hypergraphs using random sampling and contraction of hedges and hyperedges.

346