The integration of a modified balcik last mile distribution model using open road networks into a relief operations management information system

Abstract

The last mile in a disaster relief distribution chain is the delivery of goods from a central warehouse to the evacuation centers assigned for a given area. Its effectiveness relies on the proper allocation of each kind of relief good amongst the demand areas based on a given schedule. Because these operations involve a limited supply of relief goods, vehicles, and time, it is important to find ways to have more data-driven operations to satisfy as much demand as possible. There are various ways to model relief operations. One of them is Balcik's Last Mile Distribution Model, which uses linear programming to minimize routing costs as well as penalty costs for unsatisfied demands. The model provides an allocation of each kind of relief good to the demand areas visited per day. The areas visited per day would depend on the capacity of the vehicle fleet as well as on the routes that can be used. Map data used for determining routes and historical data from previous disasters are used to determine the supply and demand for relief goods while providing a benchmark for results. The study compares Balcik's Last Mile Distribution Model with other programming models intended for relief distribution to see how this is most applicable in Philippine relief scenarios. The said model is modified to fit the relief operations in the Philippines, specifically in Marikina City, specifically by changing the item types that the Balcik model would read. The model is integrated into a relief operations management information system, which will also be modified to better suit the usability needs of relief practitioners. The result is an allocation of relief goods for each evacuation area, a schedule for relief operations, as well as a visualization of the route to be used. The model provides the computational backbone for relief distribution decisions in the Philippines, allowing for more data-driven operations in the future.