Maximizing the expected net present value of a project with phase-type distributed activity durations: An efficient globally optimal solution procedure

Autor: Stefan Creemers
Přispěvatelé: Lille économie management - UMR 9221 (LEM), Université d'Artois (UA)-Université catholique de Lille (UCL)-Université de Lille-Centre National de la Recherche Scientifique (CNRS)
Jazyk: angličtina
Rok vydání: 2018
Předmět:
Schedule
Mathematical optimization
Information Systems and Management
General Computer Science
Computer science
0211 other engineering and technologies
02 engineering and technology
Type (model theory)
Management Science and Operations Research
Phase (combat)
Net present value
Industrial and Manufacturing Engineering
[SHS]Humanities and Social Sciences
SNPV
Project management
0502 economics and business
0202 electrical engineering
electronic engineering
information engineering
Duration (project management)
050210 logistics & transportation
NPV maximization
021103 operations research
Markov chain
business.industry
05 social sciences
Stochastic game
Stochastic activity durations
Schedule (project management)
Modeling and Simulation
[SHS.GESTION]Humanities and Social Sciences/Business administration
020201 artificial intelligence & image processing
Cash flow
Project scheduling
business
Zdroj: European Journal of Operational Research
European Journal of Operational Research, Elsevier, 2018, 267 (1), pp.16-22. ⟨10.1016/j.ejor.2017.11.027⟩
European Journal of Operational Research, 2018, 267 (1), pp.16-22. ⟨10.1016/j.ejor.2017.11.027⟩
ISSN: 0377-2217
1872-6860
Popis: International audience; We study projects with activities that have stochastic durations that are modeled using phase-type distributions. Intermediate cash flows are incurred during the execution of the project. Upon completion of all project activities a payoff is obtained. Because activity durations are stochastic, activity starting times cannot be defined at the start of the project. Instead, we have to rely on a policy to schedule activities during the execution of the project. The optimal policy schedules activities such that the expected net present value of the project is maximized. We determine the optimal policy using a new continuous-time Markov chain and a backward stochastic dynamic program. Although the new continuous-time Markov chain allows to drastically reduce memory requirements (when compared to existing methods), it also allows activities to be preempted; an assumption that is not always desirable. We prove, however, that it is globally optimal not to preempt activities if cash flows are incurred at the start of an activity. Moreover, this proof holds regardless of the duration distribution of the activities. A computational experiment shows that we significantly outperform current state-of-the-art procedures. On average, we improve computational efficiency by a factor of 600, and reduce memory requirements by a factor of 321.
Databáze: OpenAIRE
Externí odkaz: