This summer I had the great fortune to work with two undergraduate industrial engineering students (one from Virginia Tech, the other from the University of San Diego) on the medication distribution problem. (The students were participating in an REU site here at Institute for Systems Research.)
This is the problem of planning how to distribute medication from an RSS to multiple PODs while material continues to arrive (from the SNS and VMI) to the RSS. The goal is to find a plan that uses a fixed number of vehicles and maximizes the slack of the deliveries (as a hedge or buffer against uncertainties). We developed a two-stage routing and scheduling approach and tested it on a realistic scenario from the state of Maryland. For routing vehicles we took advantage of the TourSolver software available through the CDC. We also considered the impact of using a local distribution center as an intermediate point between the RSS and the PODs. Our technical report is now available online at http://hdl.handle.net/1903/8417.
The paper’s results for this particular scenario are not meant to prove whether LDCs are useful. The goal is to provide a methodology for creating medication distribution plans and evaluating their slack as one component in deciding which plan is best for a particular area. I would love to hear from anyone who would like to apply this approach to their situation.
