Dynamic Multi-objective Optimisation Problems with Intertemporal Dependencies
Author(s)
van Tonder, Bernard
Helbig, Mardé
Griffith University Author(s)
Year published
2020
Metadata
Show full item recordAbstract
A core goal of research in dynamic multi-objective optimisation (DMOO) is to develop algorithms that can find the best possible trade-off solutions for real-world DMOO problems (RWPs). A useful comparison of DMOO algorithms for RWPs require benchmark functions that are representative of RWPs. However, only a few standard DMOO benchmark functions contain complex intertemporal dependencies found in RWPs.
This study evaluates the performance of two DMOO algorithms on two benchmark functions (BFs) with various combinations of frequency and severity of change, as well as extended versions of these BFs that include intertemporal ...
View more >A core goal of research in dynamic multi-objective optimisation (DMOO) is to develop algorithms that can find the best possible trade-off solutions for real-world DMOO problems (RWPs). A useful comparison of DMOO algorithms for RWPs require benchmark functions that are representative of RWPs. However, only a few standard DMOO benchmark functions contain complex intertemporal dependencies found in RWPs. This study evaluates the performance of two DMOO algorithms on two benchmark functions (BFs) with various combinations of frequency and severity of change, as well as extended versions of these BFs that include intertemporal dependencies. The results indicate that the performance of the algorithms was significantly worse on the BFs with intertemporal dependencies.
View less >
View more >A core goal of research in dynamic multi-objective optimisation (DMOO) is to develop algorithms that can find the best possible trade-off solutions for real-world DMOO problems (RWPs). A useful comparison of DMOO algorithms for RWPs require benchmark functions that are representative of RWPs. However, only a few standard DMOO benchmark functions contain complex intertemporal dependencies found in RWPs. This study evaluates the performance of two DMOO algorithms on two benchmark functions (BFs) with various combinations of frequency and severity of change, as well as extended versions of these BFs that include intertemporal dependencies. The results indicate that the performance of the algorithms was significantly worse on the BFs with intertemporal dependencies.
View less >
Conference Title
Proceedings of the 2020 Genetic and Evolutionary Computation Conference (GECCO '20)
Subject
Evolutionary computation
Satisfiability and optimisation
Planning and decision making