Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WC77

4 - Analogical Evaluation of the Urban Function Combination Mode of the Integrated Transportation Hub Based on AHP Fuzzy Comprehensive Evaluation Method Siyu Tao, School of Transportation and Logistics, Southwest Jiaotong University, Chengdu, China, Feng Tao, Xinmei Chen, anjun li, Lisha Wang This paper adopts a AHP fuzzy comprehensive evaluation for the different forms of the integrated transportation hub, the AHP model is used to establish & quantify the assessed level indicators of the urban function combination modes of the integrated transportation hub, & an index evaluation system is established to evaluate the advantages & disadvantages of different combination modes accurately. Suggestions are proposed for optimization & improvement on the evaluation results, & scientific guidance are provided for the development, construction & investment operation of the urban function of the integrated transportation hub. n WC75 West Bldg 212B Online Optimization I Contributed Session Chair: Wenda Zhang, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, United States 1 - Dispatch: An Optimal Algorithm for Online Perfect Bipartite Matching with I.I.D. Arrivals Quico Spaen, University of California, Berkeley, CA, 94709, United States, Minjun Chang, Dorit Simona Hochbaum, Mark Velednitsky We introduce DISPATCH, a 0.5-competitive, randomized algorithm for the problem of weighted online perfect bipartite matching with i.i.d. arrivals. We prove that 0.5 is the best-possible competitive ratio. In this problem, we are given a known set of workers, a distribution over job types, and non-negative utility weights for each worker, job type pair. At each time step, a job is drawn i.i.d. from the distribution over job types. Upon arrival, the job must be irrevocably assigned to a worker. The goal is to maximize the expected sum of utilities after all jobs are assigned. Our work is motivated by the application of ride-hailing, where jobs represent passengers and workers represent drivers. 2 - Online Linear Programming with Production Costs Michael Fairley, PhD Candidate, Stanford University, Stanford, CA, 94305, United States, Yinyu Ye We consider a sequential decision making problem where a sequence of orders arrive and we must decide to accept or decline the order before the next order is revealed. If an order is accepted then we create a production plan to supply the products by selecting suppliers to provide each product. The offline problem is a linear program and the online problem is an online linear program where the objective coefficient and constraint matrix are revealed over time. We present an online algorithm, which learns the dual prices from the columns of the previous periods and uses the dual prices to make a decision in the current period. We present simulation results that indicate that our algorithm is near optimal. 3 - Model-based Sensor-selection Submodular Optimization: Performance Trade-offs Orlando Romero, Rensselaer Polytechnic Institute, Troy, NY, 12180, United States, Sergio Pequito Submodularity is an increasingly popular tool to address NP-hard problems defined on matroid-constrained subset cost/utility functions. Within this framework, polynomial-time approximation algorithms (such as greedy algorithms) can be implemented with theoretical worst-case guarantees. This work is motivated by sensor and actuator selection problems in the context of large-scale dynamical systems, where we show that under certain problem formulations within this context, the more well-known submodular-based bounds can be proved to be overly conservative since they fail to fully leverage the structure of the problem. 4 - A Branch-and-bound Algorithm for the One-machine Problem and Delayed Precedence Constraints Variation Wenda Zhang, University of Illinois, Urbana, IL, 61802, United States, Jason Sauppe, Sheldon H. Jacobson The one-machine scheduling problem with release and delivery times with the minimum makespan objective often arises as a subproblem for the general job shop scheduling problem. A branch-and-bound algorithm solves the problem and its variation, which allows the presence of constraints that require a delay between the completion of one job and the start of another. This paper analyzes key components of this branch-and-bound algorithm and proposes a new heuristic and a different search strategy.Computational experiments show the modified algorithm to have substantial improvement in running time and number of iterations on instances both with and without delayed precedence constraints.

n WC76 West Bldg 212C Topics for Phd Students Sponsored: Minority Issues Sponsored Session Chair: Karen T. Hicklin, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599-3260, United States 1 - Topics for PhD Students Karen T. Hicklin, University of North Carolina at Chapel Hill, B-24 Hanes Hall, Chapel Hill, NC, 27599-3260, United States This panel discussion will feature early career professionals who will share advice regarding their graduate school experiences and navigating the job market. Panelists Shannon Harris, The Ohio State University, 1262 Eastwood Ave, Columbus, OH, 43203, United States Alba Rojas-Cordova, Southern Methodist University, Dallas, TX, 75219, United States Jessye Talley, Morgan State University, 2316 Naylor Rd SE, Washington, DC, 20020, United States n WC77 West Bldg 213A Finance – Risk Management Contributed Session Chair: Shaofang Li, Southeast University, Jiangning District, Nanjing, 211189, China 1 - Computation of Optimal Conditional Expected Drawdown Portfolios Alex Papanicolaou, University of California-Berkeley, Mountain View, CA, 94041, United States, Lisa R. Goldberg, Ola Mahmoud We introduce a partial differential equation approach to computing and minimizing the risk measure of Conditional Expected Drawdown (CED) of Goldberg and Mahmoud (2016). We derive a methodology that allow ones to perform function and gradient evaluation for a first-order optimization routine to a convex portfolio optimization problem for computing minimum CED portfolios. We apply this to a log-normal pricing model and show how under this framework, minimum drawdown portfolios relate to minimum variance and minimum shortfall portfolios. 2 - Merging for Stabilizing Financial Networks Aein Khabazian, University of Houston, 4796 Cullen Blvd, Houston, TX, 77004, United States, Markku Kallio, Jiming Peng We analyze the possibility of the mergers of solvent and in-solvent banks to stabilize a financial network based on the extended Eisenberg-Noe’s model where the liquidation cost is considered and the liabilities have different seniorities. We develop a model to identify the best match in the system such that the total gain obtained through the merger is maximized. A subsidized merger where the social planner provides some bail-outs to cover part of the liabilities of the insolvent bank is also considered. We use a network of major European banks to evaluate the performance of the pure merger and subsidized merger in the system. 3 - Banking Sector Reform, Competition, and Bank Stability Shaofang Li, Southeast University, No.70, Suyuan Avenue, Jiangning District, Nanjing, 211189, China, Wenping Wang This study tests the impact of banking sector reform and competition on bank stability in 14 transition countries in 1998-2016. The results find negative relationship between the market power and bank stability, but positive relationship for bank reform. The results also suggest that the positive relationship between bank competition and stability is more pronounced if banking sector experienced higher financial development and has better financial information transparency. The effect of competition in reducing bank risk is less pronounced if the banks are exposed to higher activity restrictions and grater stringency on capital requirements, particularly in more competitive markets.

495