Two-level, two-objective evolutionary algorithms for solving unit commitment problems

Georgopoulou, CA; Giannakoglou, KC

dc.contributor.author	Georgopoulou, CA	en
dc.contributor.author	Giannakoglou, KC	en
dc.date.accessioned	2014-03-01T01:32:14Z
dc.date.available	2014-03-01T01:32:14Z
dc.date.issued	2009	en
dc.identifier.issn	0306-2619	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/20094
dc.subject	Unit commitment	en
dc.subject	Evolutionary algorithms	en
dc.subject	Multilevel search	en
dc.subject	Multiobjective optimization	en
dc.subject	Stochastic power demand distribution	en
dc.subject.classification	Energy & Fuels	en
dc.subject.classification	Engineering, Chemical	en
dc.subject.other	GENETIC ALGORITHM	en
dc.subject.other	LAGRANGIAN-RELAXATION	en
dc.subject.other	ECONOMIC-DISPATCH	en
dc.subject.other	OPTIMIZATION	en
dc.subject.other	NETWORKS	en
dc.subject.other	SEARCH	en
dc.subject.other	SYSTEM	en
dc.title	Two-level, two-objective evolutionary algorithms for solving unit commitment problems	en
heal.type	journalArticle	en
heal.identifier.primary	10.1016/j.apenergy.2008.08.001	en
heal.identifier.secondary	http://dx.doi.org/10.1016/j.apenergy.2008.08.001	en
heal.language	English	en
heal.publicationDate	2009	en
heal.abstract	A two-level, two-objective optimization scheme based on evolutionary algorithms (EAs) is proposed for solving power generating Unit Commitment (UC) problems by considering stochastic power demand variations. Apart from the total operating cost to cover a known power demand distribution over the scheduling horizon, which is the first objective, the risk of not fulfilling possible demand variations forms the second objective to be minimized. For this kind of problems with a high number of decision variables, conventional EAs become inefficient optimization tools, since they require a high number of evaluations before reaching the optimal solution(s). To considerably reduce the computational burden, a two-level algorithm is proposed. At the low level, a coarsened UC problem is defined and solved using EAs to locate promising solutions at low cost: a strategy for coarsening the UC problem is proposed. Promising solutions migrate upwards to be injected into the high level EA population for further refinement. In addition, at the high level, the scheduling horizon is partitioned in a small number of subperiods of time which are optimized iteratively using EAs, based on objective function(s) penalized to ensure smooth transition from/to the adjacent subperiods. Handling shorter chromosomes due to partitioning increases method's efficiency despite the need for iterating. The proposed two-level method and conventional EAs are compared on representative test problems. (C) 2008 Elsevier Ltd. All rights reserved.	en
heal.publisher	ELSEVIER SCI LTD	en
heal.journalName	APPLIED ENERGY	en
dc.identifier.doi	10.1016/j.apenergy.2008.08.001	en
dc.identifier.isi	ISI:000265033400029	en
dc.identifier.volume	86	en
dc.identifier.issue	7-8	en
dc.identifier.spage	1229	en
dc.identifier.epage	1239	en