Efficient multiple objective optimal control of dynamic systems with integer controls

In practical optimal control problems both integer control variables and multiple objectives can be present. The current paper proposes a generic and efficient solution strategy for these multiple objective mixed-integer optimal control problems (MO-MIOCPs) based on deterministic approaches. Hereto,...

Full description

Saved in:
Bibliographic Details
Main Authors: Logist, Filip (Author) , Sager, Sebastian (Author) , Kirches, Christian (Author) , Van Impe, Jan F. M. (Author)
Format: Article (Journal)
Language:English
Published: 23 May 2010
In: Journal of process control
Year: 2010, Volume: 20, Issue: 7, Pages: 810-822
ISSN:0959-1524
DOI:10.1016/j.jprocont.2010.04.009
Online Access:Resolving-System, lizenzpflichtig, Volltext: https://doi.org/10.1016/j.jprocont.2010.04.009
Verlag, lizenzpflichtig, Volltext: https://www.sciencedirect.com/science/article/pii/S0959152410000806
Get full text
Author Notes:F. Logist, S. Sager, C. Kirches, J.F. Van Impe
Description
Summary:In practical optimal control problems both integer control variables and multiple objectives can be present. The current paper proposes a generic and efficient solution strategy for these multiple objective mixed-integer optimal control problems (MO-MIOCPs) based on deterministic approaches. Hereto, alternative scalar multiple objective optimisation techniques as normal boundary intersection and normalised normal constraint are used to convert the original problem into a series of parametric single objective optimisation problems. These single objective mixed-integer optimal control problems are then efficiently solved through direct multiple shooting techniques which exploit convex relaxations of the original problem. Moreover, these relaxations enable to quickly approximate the final solution to any desired accuracy (without the need of solving integer problems). Consequently, the set of Pareto optimal solutions of the MO-MIOCP can be accurately obtained in highly competitive computation times. The proposed method is illustrated on (i) a testdrive case study with a complex car model which includes different gears and conflicting minimum time-minimum fuel consumption objectives, and (ii) a jacketed tubular reactor case study with conflicting conversion, heat recovery and installation costs.
Item Description:Gesehen am 29.03.2023
Physical Description:Online Resource
ISSN:0959-1524
DOI:10.1016/j.jprocont.2010.04.009