Automatiseeritud ressursside optimiseerimine äriprotsessides
Files
Date
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Tartu Ülikool
Abstract
Äriprotsesside optimiseerimine on aeganõudev tegevus. Seda põhjustab võimalike optimiseeringute suur arv. Antud olukord sunnib analüütikuid probleemi lahendama katse-eksitus meetodi abil ning parima võimaliku lahenduse leidmine pole garanteeritud. Automatiseeritud äriprotsesside parandamist on uuritud ning parimaid tulemusi on saavutatud geneetiliste algoritmide abil ressursside jaotuse muutmisel. See protsess on väga aeganõudev ning tihtipeale suudetakse optimaliseerida ainult ühe tulemuslikkuse näitaja suhtes. Eelnevalt nimetatud probleemide lahendamiseks on käesoleva magistritöö eesmärgiks olnud arendada automaatne äriprotsessi ressursside jaotuse optimaliseerija, mis suudab protsessi parandada korraga mitme tulemuslikkuse näitaja suhtes. Tähtsat rolli omas ka optimaliseerija kiirus. Üldised tulemused olid positiivsed ning tõestasid juhtumiuuringu abil, et matemaatilist optimaliseerimistehnikat nimetusega hill-climbing on võimalik rakendada äriprotsessi resursside jaotamise probleemi lahendamisel.
Business process optimization is a time-consuming activity. Mainly because the amount of possible modifications is very large and a business analyst usually approaches the task with a trial-and-error method. As a result the analyst is not guaranteed to find the best possible solution. Currently some automated process optimizers do exist but their functionality is limited and the processing time is long. In this thesis I implemented an automated resource allocation optimizer based on hill-climbing. The optimizer can optimize on execution time, cost or combination of thereof when given certain resource constraints, expressed in terms of minimum and maximum size for each resource pool. The overall results were positive and on a basis of a case study it was proven that a hill-climbing algorithm can be applied to a resource allocation problem. In a reasonable time frame the optimizer was capable of finding the best cycle time and cost trade-off curve in two-dimensional space.
Business process optimization is a time-consuming activity. Mainly because the amount of possible modifications is very large and a business analyst usually approaches the task with a trial-and-error method. As a result the analyst is not guaranteed to find the best possible solution. Currently some automated process optimizers do exist but their functionality is limited and the processing time is long. In this thesis I implemented an automated resource allocation optimizer based on hill-climbing. The optimizer can optimize on execution time, cost or combination of thereof when given certain resource constraints, expressed in terms of minimum and maximum size for each resource pool. The overall results were positive and on a basis of a case study it was proven that a hill-climbing algorithm can be applied to a resource allocation problem. In a reasonable time frame the optimizer was capable of finding the best cycle time and cost trade-off curve in two-dimensional space.