Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SA11

3 - Using Second-order Information to Accelerate Incremental Gradient Methods Hoi-To Wai, Arizona State University, Tempe, AZ, United States, Wei Shi, Cesar A. Uribe, Angelia Nedich, Anna Scaglione Finite sum problems arise in applications of machine learning and control systems. To handle them, one may apply incremental method that divides the objective into chunks, processed one-at-a-time. Existing methods suffer from slow convergence limited by number of chunks involved. This talk introduces a curvature aided technique for incremental gradient methods. We propose a CIAG method using the technique and show that its asymptotic convergence rate is independent of the number of chunks. Two extensions are discussed: Nesterov acceleration is applied and an accelerated rate can be achieved; for distributed optimization, we use a random-walk based method with a rate robust to graph topology. 4 - Projective Splitting with Forward Steps: Asynchronous and Block-iterative Operator Splitting Patrick Johnstone, Rutgers, 100 Rockafeller Rd., Piscataway, NJ, 08854, United States, Jonathan Eckstein For a recently proposed projective splitting framework, we show how to replace the fundamental subproblem calculation using a backward step with one based on two forward steps. The resulting algorithms have the same kind of coordination procedure and can be implemented in the same block-iterative, distributed, and asynchronous manner, but may perform backward steps on some operators and forward steps on others. Prior algorithms in the projective splitting family have used only backward steps. Forward steps can be used for Lipschitz- continuous operators if the stepsize is bounded by the inverse of the Lipschitz constant, or if a backtracking procedure is used. We also derive convergence rates. n SA10 North Bldg 125A Joint Session MSOM/Practice Curated:Empirical and Analytical Research to Improve Humanitarian Service Delivery Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Mahyar Eftekhar, Arizona State University, Tempe, AZ, 85287, United States 1 - What Drives Humanitarian Organizations’ Donation Income? An Empirical Investigation Iman Parsa, Arizona State University, Mahyar Eftekhar, Charles J. Corbett For their survival, humanitarian organizations greatly depend on donations they receive. We empirically study how different factors impact individual donations and government grants that medium and large humanitarian organizations ?receive to see whether, and how, they can adopt policies to attract more donations. These factors include elements related to organizations’ internal operations, such as program spending ratio and transparency, as well as external factors, like media exposure. 2 - A Field Study of Charitable Giving Reveals that Reciprocity Decays over Time Amanda Chuan, University of Pennsylvania, Philadelphia, PA, United States, Judd Kessler, Katherine MIlkman We examine how reciprocity changes over time using a large field study. We analyze data from a hospital system on 18,000 donation requests to former patients in the four months following their first hospital visit. We exploit quasi- experimental variation in solicitation timing and find that an extra 30-day delay between the provision of medical care and a donation solicitation decreases the likelihood of a donation by 30%. Our findings have important implications for models of economic behavior, which currently fail to incorporate reciprocity’s sensitivity to time. The fact that reciprocal behavior decays rapidly also suggests the importance of capitalizing quickly on quid pro quo opportunities. 3 - Communities in the Crossfire: How Companies Can Do Well by Doing Good Andres Fernando Jola-Sanchez, Texas A&M University, College Station, TX, United States, Alfonso J. Pedraza-Martinez We study how social investments in conflict zones affect firms’ operational performance. Social investments can improve the well-being of communities in war-torn areas, but how do firms benefit from making these investments? We collect data from all the oil firms in Colombia and examine a law that compelled a group of these firms to spend 1% of their budget on social investmentsùthe Colombian civil war is rampant in oil regions. We find this group obtains higher operating margins compared to the group the law did not affect. By influencing workforce, sourcing and logistics processes, social investments reduce the frequency of disruptions affecting firms in conflict zones.?

4 - Supply Management for Rapid-onset Disasters: Effects of Supply, Demand, and Budget Uncertainty Mahyar Eftekhar, Arizona State University, BA 433, Main Campus, P.O. Box 874706, Tempe, AZ, 85287, United States, Jing-Sheng Jeannette Song, Scott Webster To fulfill beneficiaries’ demands, humanitarian organizations should design a cost- efficient and time-effective procurement policy. We consider and analyze two common supply management policies: pre-positioning and local-purchasing. Our analysis takes demand, supply, and budget uncertainties into account. 5 - Do Optimization Models for Humanitarian Operations Need a Paradigm Shift Harwin de Vries, INSEAD, Boulevard de Constance, Fontainebleau, 77210, France, Luk N. Van Wassenhove Optimization approaches for planning and routing of humanitarian field operations have been studied a lot. Yet, their adoption in practice seems absent. Based on interviews, literature, and modeling results, we discuss the applicability and cost-effectiveness of such approaches and identify areas where future research is needed. Supply Chain Finance/Risk Management Sponsored: Manufacturing & Service Oper Mgmt/iFORM Sponsored Session Chair: Heikki Peura, Imperial College Business School, London, SW7 2AZ, United Kingdom 1 - Trade Credit Provision and Inventory Performance Christopher J. Chen, London Business School, London Business School, Regent’s Park, London, NW14SA, United Kingdom, Nitish Jain, S. Alex Yang Long payment terms in the form of trade credit are believed to be harmful to suppliers and was a driving force behind French regulation in 2008 limiting its duration. Using data on European retailers, we exploit variation in firms’ exposure to the law in order to quantify its impact on trade credit usage. Leveraging this natural experiment further, we test theories on the relationship between trade credit and operational decisions. 2 - Supplier Diversification Under Buyer Risk Nikos Trichakis, MIT, Jiri Chod, Gerry Tsoukalas When should a firm diversify its supply base? Most extant theories attribute supplier diversification to supplier risk. We develop a new theory that attributes diversification to buyer risk. When suppliers are subject to buyer default risk, buyers may take costly action to signal creditworthiness so as to obtain more favorable terms. But once signaling costs are sunk, buyers sourcing from a single supplier become vulnerable to future holdup. Although ex ante supply base diversification can be effective at alleviating this problem, we show that it comes at the expense of higher upfront signaling costs. We resolve the ensuing trade-off and show that diversification emerges as the preferred strategy. 3 - The Value of Supply Chain Disruption Duration Information Bill Schmidt, Cornell University, Ithaca, NY, United States, Mili Mehrotra Using the supply chain and production data from a multinational division of a Fortune 500 manufacturing firm, we quantify the operational performance impact of disruption duration information. We identify several factors that influence the value of information, helping the firm to eliminate anywhere from less than 1% to as much as 100% of the cost of the disruption. n SA11 North Bldg 125B

5