2015 Informs Annual Meeting

WA27

INFORMS Philadelphia – 2015

WA28 28-Room 405, Marriott Bidding Mechanisms Cluster: Auctions Invited Session

5 - Open Priority Scheduling Protocol for Sourcing and Supply Management Katariina Kemppainen, School of Business, Aalto University, Runeberginkatu 22-24, Helsinki, 00076 Aalt, Finland, katariina.kemppainen@aalto.fi, Ari Vepsäläinen Are you happy with the first-come-first-served priority rule when ordering goods online? The extensive scheduling literature has few answers from real-life applications, whereas in procurement case studies abound but theoretical analyses are few. We propose an open protocol for manufacturing and procurement which beats consistently not only FCFS but any other rule. It offers a sophisticated solution for one universal rule. Our protocol is backed up with simulation results and real-life cases. WA27 27-Room 404, Marriott Multi-objective Optimization and Applications Sponsor: Multiple Criteria Decision Making Sponsored Session Chair: Matthias Ehrgott, Professor, Department: Management Science, Lancaster University, The Management School, Lancaster, 00, LA1 4YX, United Kingdom, m.ehrgott@lancaster.ac.uk 1 - Ultra-high-dimensional Optimization for Trade Space Exploration: Challenges and Lessons Learned Matthew Hoffman, Sandia National Laboratories, P.O. Box 5800 MS 1188, Albuquerque, NM, 87185-1188, United States of America, mjhoffm@sandia.gov, Alexander Dessanti, Stephen Henry, Jack Gauthier In projects with many conflicting stakeholder objectives, negotiating compromise requires understanding tradeoffs. Existing multiobjective approaches build “coalitions” of objectives to reduce dimensionality – either explicitly via aggregation, or implicitly – which favors solutions that compromise across many objectives, but can severely obfuscate or distort the tradeoffs between them. We discuss our progress thus far in true ultra-high-dimensional optimization for trade space exploration. 2 - Special Constraint Treatment for Multi-objective Optimization Sijie Liu, Graduate Student, University of Alabama, 714 1/2 12th St. B, Tuscaloosa, AL, 35401, United States of America, sliu28@crimson.ua.edu Several methods has been developed for finding pareto set and pareto front for multi-linear objective optimization problem subjective to one additional special constraint(such as bi-linear constraint or N-linear constraint) over a bounded interval domain subjective to linear equality constraint. These methods provide us fast approaches to plot the entire pareto frontier. Algorithms have been proved and numerical results demonstrate the effectiveness of these methods. 3 - A Nonparametric Approach to the Multi-Objective Sequential Decision Problem Many choices are presented one at a time. You must decide whether to choose the current choice and stop the search process or to reject it and move to the next stage. The decision is irrevocable and each choice is evaluated based on multiple objectives. We propose a rank-based optimal decision strategy that minimizes the weighted rank of the selected choice. It can be shown that many sequential decision problems are special cases of the generalized multi-objective sequential decision problem. 4 - Deliverable Radiotherapy using Multiobjective Optimisatio and Column Generation Matthias Ehrgott, Professor, Department: Management Science, Lancaster University, The Management School, Lancaster, 00, LA1 4YX, United Kingdom, m.ehrgott@lancaster.ac.uk, Kuan-min Lin We propose a column generation based approach to compute deliverable radiotherapy treatment plans based on a multi-objective optimisation problem. We compute a representative set of non-dominated treatment plans using the revised normal boundary intersection method. The generation of each non- dominated treatment plan uses column generation to generate apertures that directly deliverable, obviating the need for a separate segmentation step that deteriorates plan quality. Young H. Chun, Professor, Louisiana State University, E. J. Ourso College of Business, Baton Rouge, LA, 70820, United States of America, prof@drchun.net

Chair: Bernardo Quiroga, Assistant Professor, Business and Behavioral Science, Clemson University, 100 Sirrine Hall, Clemson, SC, 29634, United States of America, bfquirog@gmail.com 1 - One-dimensional Strategyproof Facility Location Itai Feigenbaum, Columbia University, New York, NY, United States of American, itai@ieor.columbia.edu, Chun Ye, Jay Sethuraman Consider a set of agents on an interval, where a planner wishes to locate a facility so as to maximize some social benefit function. The agents have linear preferences over the location of the facility, and their locations are unknown to the planner. Thus, the planner wishes to locate the facility in a strategyproof manner while approximating social benefit. We discuss mechanisms, lower bounds, and characterizations for various versions of this model. 2 - Greening Multi-tenant Data Center Demand Response with Supply Function Bidding Niangjun Chen, California Institute of Technology, Pasadena, CA, United States of American, ncchen@caltech.edu, Adam Wierman, Shaolei Ren, Xiaoqi Ren Data centers have become critical resources for emergency demand response (EDR). In this talk, we focus on “greening” demand response in multi-tenant data centers by incentivizing tenants’ load reduction and reducing on-site diesel generation. Our proposed mechanism, ColoEDR, which is based on parameterized supply function mechanism, provides provably near-optimal efficiency guarantees, both when tenants are price-taking and when they are price- anticipating. 3 - Optimal Bidding for Bundles in Sequential Auctions Karti Puranam, Assistant Professor, La Salle University, 1900 W Olney Ave, Philadelphia, PA, 19141, United States of America, puranam@lasalle.edu, Michael Katehakis We study the problem of optimal bidding for a firm that in each period procures items to meet a random demand by participating in a finite sequence of auctions where in each auction involves bids for bundles of items. We develop a new model for a firm where its item valuation derives from the sale of the acquired. We establish monotonicity properties for the value function and the optimal dynamic bid strategy and we present computations. 4 - Complexity and Transparency in Sealed-bid Procurement Auctions with Multi-dimensional Bids Bernardo Quiroga, Assistant Professor, Business and Behavioral Science, Clemson University, 100 Sirrine Hall, Clemson, SC, 29634, United States of America, bfquirog@gmail.com, Brent Moritz, V. Daniel R. Guide, Jr. We experimentally analyze A+B (price-and-quality) bidding behavior in two closely related sealed-bid scenarios: One where the rule to assign the contract is transparently communicated to bidders before they submit their offers, and another where the assignment rule is only known to the buyer and not to the bidders. Our results show substantial losses in terms of social welfare and buyer surplus as a direct effect of transparency loss in this procurement system.

WA29 29-Room 406, Marriott Applied Analytics Sponsor: Analytics Sponsored Session

Chair: Jon Alt, Assistant Professor, Naval Postgraduate School, Department of Operations Research, Naval Postgraduate School, Monterey, CA, 93943, United States of America, jkalt@nps.edu 1 - Improved U.S. Army Reserve Stationing using Recruitable Market Demographics Nathan Parker, U.S. Army, 4120 Crest Rd, Pebble Beach, CA, 93953, United States of America, nparker@nps.edu The objective of this work is to develop a model to predict a U.S. Army Reserve (USAR) unit’s manning level based on the demographics of the unit’s reserve center recruitable market (RCRM). This study first develops an allocation method to determine the RCRM available to each reserve center. Classification and regression models are then developed to determine the ability of the RCRM to support the reserve center’s manning requirements.

382