Satellite imaging order scheduling with stochastic weather condition forecast

This paper deals with the imaging order scheduling problems of FORMOSA-2, a low-earth-orbit, remote-sensing satellite which takes images of ocean and landmass in the vicinity of Taiwan according to customers' requests. The satellite imaging order scheduling considers current and future weather conditions to provide an imaging plan to satisfy customer requirements with quality imaging pictures. We first formulate the problem into a stochastic integer programming problem and adopt the rolling horizon approach to solve for this problem. A nominal plan is generated according to the predetermined or forecast conditions. This nominal plan is subject to be adjusted with the updated information updated during its execution. The Lagrangian relaxation technique is used to solve for the nominal imaging plan. Numerical results indicate that this approach is effective to generate near-optimal order plans.