Using Linear Programming to Design Oil Pollution Detection Schedules

Abstract

Abstract This paper presents a flight scheduling system for sensor (infra-red and ultra-violet) equipped aircraft whose mission is the detection and prevention of harbor and coastal oil and hazardous material pollution. The aircraft scheduling model is a form of the stochastic traveling salesman problem which can be solved using linear programming. The objective of the model is to maximize the expected number of pollution incidents detected per pollution flight. The model requires as input parameters representing probabilities of pollution incidents occurring for different geographical sectors. These parameters are estimated using coastal and harbor shipping statistics for petroleum and hazardous material. The shipping statistics considered include movement of petroleum products along the Atlantic, Pacific and Gulf coasts as well as the movement of such commodities in the intra-coastal waterway system and the Great Lakes. Particular port characteristics and past pollution incident statistics are also used to estimate required parameters of the scheduling model. The constraints in the model concern the availability of aircraft, flight hours, mission hours, and the significant environmental factors which affect the performance of the sensors.