Article ID: | iaor1998792 |
Country: | United States |
Volume: | 43 |
Issue: | 3 |
Start Page Number: | 386 |
End Page Number: | 402 |
Publication Date: | Mar 1997 |
Journal: | Management Science |
Authors: | Wilhelm W.E., Srinivasa A.V. |
Keywords: | geography & environment, heuristics, programming: integer |
The Tactical Decision Problem (TDP) associated with oil spill clean up operations prescribes the time-phased allocation of available components over the planning horizon so that the clean up requirement at each critical time point is met. The objective is to minimize response time to allow for the most effective clean up possible. In this paper, we formulate the TDP as a general integer program. We devise a method based on graph theory to efficiently generate response system types (RSTs), including constituent components, the locations from which each of the component types is obtained, and the staging area at which that RST is composed. We present several pre-processing methods and derive expressions for bounds on decision variables to facilitate solution. We then develop two heuristics to obtain approximate solutions to the TDP. The first heuristic is an LP-based method, while the second uses a combination of LP relaxation and branch and bound. The two heuristics are compared on problems that are based on realistic scenarios representing application in the Galveston Bay Area.