The multi-satellite image acquisition schedulingproblem through MCDM and an extended longestpath algorithm

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

The problem of image acquisition scheduling frommultiple earth-observation satellites is ultimately a complexdecision-making problem comprised of an intrinsic valuation andscheduling of imaging requests which correspondingly produce aschedule that follows temporal and operational constraints andinclude the most highly preferred requests. Extensive researchcan be found in the literature on the scheduling part of theproblem, but the matter of defining and understanding thepreference structure to support and improve the decision makingof the satellite operator (SO) has been widely neglected. In thisstudy, an innovative three-step solution method is developed,which firstly consists of a multi-criteria scoring approach utilizingthe ELECTRE-III method to produce a valuation based onpairwise evaluations of each feasible imaging attempt, secondly, agraph-based longest path heuristic introducing the characteristicsof interdependent and allowed nodes for scheduling, and thirdly,a weight space analysis (WSA) of the scoring approach to supportthe SOs in obtaining a weight setting that best reflects thecommon goal of the conflicting objectives within the satelliteoperation. A comparative analysis is performed to provideinsight into the quality of the solution approach by associatingpreferences of SOs in the industry. The longest path heuristicseems to be within 10% performance of an exact schedulingapproach while being significantly faster and more capable ofhandling larger problems scenarios. For the satellite operations,the ELECTRE-III scoring approach and the WSA provides aframework that increases the explainability, insight and abilityto introduce preferences into the decision process.
Original languageEnglish
JournalIEEE Transactions on Systems, Man, and Cybernetics: Systems
ISSN1083-4427
Publication statusSubmitted - 2 Nov 2020

Cite this