Thursday, December 26, 2013

Operations Research in Disruption Management May Have Helped UPS and FedEx with Timely Deliveries for Christmas


Many of us have enjoyed celebrating the special holidays this time of the year and  are recovering from all the wonderful and, yet, frenetic activities surrounding the shopping, decorating, cooking, visiting, and partying!


Not so for those who work for our logistics companies such as UPS and FedEx whose trucks you may have seen driving through your neighborhoods many times and even late in the evening this past week to deliver packages for Christmas, which was yesterday.

You may have also caught the news of the package shipment delays this season and were anxiously awaiting the package that you ordered for a relative, neighbor, or friend, which still has not arrived.

Timely deliveries are extremely important, especially when there is a big day such as Christmas. In fact, time is a strategic advantage, as important as cost and even quality, which we have argued in our paper:
A Supply Chain Network Game Theoretic Framework for Time-Based Competition with Transportation Costs and Product Differentiation,
Anna Nagurney and Min Yu, to appear in Optimization in Science and Engineering - In Honor of the 60th Birthday of Panos M. Pardalos, edited by S. Butenko, C. A. Floudas, and Th. M. Rassias, Springer, New York, 2014.

UPS, in a statement, reported on CNN.com, explained that "the volume of air packages in our system exceeded the capacity of our network immediately preceding Christmas so some shipments were delayed." "We know how hard it is for everyone to receive their holiday packages, and we're working around the clock to resolve this issue," UPS spokeswoman Natalie Black said.

I am sure that many out there who work (and even teach)  in operations and logistics and conduct research on this great subject had an "Aha!" moment. This sounded like the maximal flow problem in operations research, which is a classical problem, but which needed to include stochastic elements associated with possible disruptions. According to Black, UPS underestimated the volume of packages and the previous severe weather in the Dallas area had already created a backlog. Then came "excess holiday volume" during a compressed time frame, since the period between Thanksgiving and Christmas was shorter than usual this year. (Also, UPS gives their employees Christmas off, which they surely need.) This issue lit up social media on the Internet around the globe.

FedEX also apologized for some delays and noted that the company handled 275 million shipments this year between Thanksgiving and ChristmasSome FedEx custimers were able to pick up their packages at local FedEx centers that were open on Christmas day.  "We're sorry that there could be delays and we're contacting affected customers who have shipments available for pickup," said Scott Fiedler, a spokesman for FedEx Corp.

Those of us who work in disaster and disruption management (and soon I will be teaching again my Humanitarian Logistics and Healthcare course at the Isenberg School of Management) know that there are many synergies and crossovers between corporate logistics and humanitarian logistics and some of the best practices in one space can be adapted to the other.

In a special issue of the journal Transportation Research A on Network Vulnerability in Large-Scale Transport Networks, our article:  A Bi-Criteria Indicator to Assess Supply Chain Network Performance for Critical Needs Under Capacity and Demand Disruptions,
Qiang Qiang and Anna Nagurney,  appeared in volume 46(5): (2012) pp 801-812.  In this paper, we developed a supply chain/logistics network model for critical needs in the case of disruptions. The objective is to minimize the total network costs, which are  generalized costs that may include the monetary, risk, time, and social costs.  The model assumes that disruptions may have an impact on both the network link capacities as well as on the product demands. Two different cases of disruption scenarios are considered. In the first case, we assume that the impacts of the disruptions are mild and that the demands can be met.  In the second case, the demands cannot all be satisfied. For these two cases, we propose two individual performance measures. We then construct a bi-criteria measure to assess the supply chain network performance. An algorithm is described which is applied to solve a spectrum of numerical examples in order to illustrate the new concepts.

This bi-criteria measure  considers the following factors:
  • Supply chain capacities may be affected by disruptions;
  • Demands may be affected by disruptions; and
  • Disruption scenarios are categorized into two types.
And, in order to determine whether the demands can be satisfied under a specific disruption scenario (which, of course, can include weather or another natural disaster)  we first solve the maximal flow problem, which is a classical network optimization problem in operations research.

Both UPS and FedEx perform, in general, amazing work and take advantage of advanced analytics but as this holiday season revealed there is more that can be done! As the incoming President of INFORMS, Dr. Stephen Robinson, alluded in an interview, sometimes it takes many years for research to make it into practice. I think that this needs to  change.