Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MA04

2 - An Analysis of Product Innovation as Recombinant Search using Topic Modelling Philipp Cornelius, Rotterdam School of Management, Rotterdam, Netherlands Product innovation is often understood as the recombination of componentsùsuch as ideas, knowledge, input factors, and technologies. An important question is whether certain recombination patterns increase innovation success. The existing literature has predominantly studied recombination of technologies using patent data. I extend this literature by studying recombination during the development of new consumer products. Interestingly, recombination strategies that create successful technologies (such as patents) do not always coincide with product success. 3 - Revenue Management in Crowdfunding Jiding Zhang, The Wharton School, 3730 Walnut St, 500 Jon M. Huntsman Hall, Philadelphia, PA, 19104, United States, Sergei Savin, Senthil Veeraraghavan Crowdfunding, a mechanism in which funds are raised online using small donations from a large number of individual donors, has recently emerged as a popular approach to funding new ideas. In our paper, we model a setting where a creator of a crowdfunding project selects the amount of contribution it requests from donors and the duration of crowdfunding campaign with the goal of maximizing the raised amount. Our analysis provides project creators with detailed, practical, and intuitive guidelines on how to successfully manage the n MA04 North Bldg 122A Algorithmic and Polyhedral Analysis of Applied MIP Problems Sponsored: Optimization/Integer and Discrete Optimization Sponsored Session Chair: Manish Bansal, Virginia Tech., Blacksburg, VA, 24060, United States 1 - Recent Advances on Solving Generalizations of the Single Item Capacitated Lot-sizing Problem Kartik Kulkarni, Virginia Tech, Manish Bansal In this talk, we present our recent advances on solving a generalization of the classical single-item economic lot-sizing problem where the total production capacity in each period can be the summation of some binary multiples of several capacity modules of different sizes. We also develop a new algorithm for the lot- sizing problem with piecewise concave production costs and concave holding costs. Finally, we present results of our computational experiments performed on the aforementioned problems. 2 - Multi-module Capacitated Network Design Problem and Its Generalizations Haochen Luo, Texas A&M University, College Station, TX, United States, Kiavash Kianfar In this work, we consider the multi-module capacitated network design (MMND) problem, where minimum cost flows and flow transfer capacities are to be decided on a network and capacities are composed of multiples of differently sized modules. We also study generalizations of MMND to networks with special requirements so that more complicated decisions need to be made. For both MMND and its generalizations, we introduce new classes of facet-defining valid inequalities not obtainable from polyhedral results in previous research. We present computational results that show the effectiveness of the new classes of inequalities for MMND and its generalizations. 3 - On the Robust Single Machine Scheduling Problem with Relative Robust Deviation Juliana Arango, Clemson University, 211 Fernow St, Clemson, SC, 29634, United States, J. Cole Smith We consider a single machine scheduling environment where the performance criterion is the total flow time, and processing times are uncertain but belong to a distinct interval for each job. We consider the case in which the uncertainty level in the problem is limited by enforcing the condition that only a constant number of jobs can have a processing time greater than the interval’s lower bound. We examine characteristics of the problem for maximizing regret given a job processing sequence, and subsequently explore algorithms for minimizing maximum regret. revenue generation process in a crowdfunding campaign. 4 - Learning as a Signal in Online Debt Market Qiang Gao, Baruch College, City University of New York, 55 Lexington Ave, New York, NY, 10010, United States Abstract not available.

4 - Remote Microgrid Design Optimization under Photovoltaic and Load Uncertainty Alexander Zolan, University of Texas-Austin, Austin, TX, United States, David Morton, Michael S. Scioletti, Amanda Hering We consider a microgrid that consists of a battery, photovoltaic (PV) systems, and diesel generators, and optimize the components it comprises and a corresponding dispatch strategy at hourly fidelity to minimize procurement, operations and maintenance, and fuel costs. We use historical weather data to generate base load and PV system power output scenarios. We present a case study with weather profiles from different locations and show that solutions to instances that account for these uncertainties produce a lower-cost, more robust solution than solutions which use point forecasts as input. n MA05 North Bldg 122B Large-Scale Optimization Algorithms Sponsored: Optimization/Linear and Conic Optimization Sponsored Session Chair: Yuyuan Ouyang, Clemson University, Clemson, SC, United States 1 - Quadratic Averaging for Smooth Convex Optimization Michael Eldredge, Clemson University, Clemson, SC First-order methods have been a popular research area in recent years, due to their efficiency on solving big data optimization problems. There had been many studies on designing efficient gradient based algorithms and understanding their performance. In this talk, we consider the interpretation of a class of accelerated gradient methods, by extending recent studies on quadratic averaging and estimating sequences. Based on the fact that many gradient based algorithms can be interpreted as designs of sequences of quadratic approximation, we design a quadratic averaging framework for solving smooth convex optimization with optimal performance. 2 - A Modified Conditional Gradient Sliding Method for Structured Convex Optimization Hamid Nazari, Clemson University, Clemson, SC, 29631, United States, Yuyuan Ouyang Based on the conditional gradient sliding (CGS) method proposed by Lan and Zhou, we design a modified CGS method for structured convex optimization. Similar as the CGS method, when minimizing structured convex functions, the proposed method is able to skip some time-consuming operations from time to time. Our modification is mainly on the acceleration technique of the sliding scheme. We prove that the convergence rate of the proposed method is at least as good as the CGS method. Some preliminary computational results comparing the proposed method and CGS will also be presented. 3 - Inexact Augmented Lagrangian Method for Convex Optimization Yangyang Xu, Rensselaer Polytechnic Institute, Department of Mathematical Sciences, 110 8th Street, Troy, NY, 12118, United States, In this talk, I will present inexact ALMs for convex programs with both equality and inequality constraints. For these problems, we establish the global convergence rate of inexact ALM and estimate its iteration complexity in terms of the number of gradient evaluations. We first establish an ergodic convergence rate result. By relating to the inexact proximal point algorithm, we also show a nonergodic convergence rate result of inexact ALM that uses geometrically increasing penalty parameters. The nonergodic iteration complexity result is in the same order as that for the ergodic result. Numerical tests on QCQP are conducted to compare the performance of the inexact ALM with different settings. 4 - Random Gradient Extrapolation for Distributed and Stochastic Optimization Yi Zhou, ISyE Ga Tech, 755 Ferst Drive, NW, Atlanta, GA, 30332, United States, Guanghui Lan In this talk, we consider a class of finite-sum convex optimization problems defined over a multiagent network with $m$ agents connected to a central server. Our major contribution is to develop the random gradient extrapolation method (RGEM), which does not require any exact gradient evaluation even for the initial point, but achieve the optimal linear rate in terms of the number of gradient evaluations. For stochastic case, RGEM maintains the optimal sublinear rate in terms of the number of stochastic gradient computations, but attains linear communication complexity. This is the first time that these complexity bounds have been obtained for distributed and stochastic optimization problems.

120