Back to Search
Start Over
Modeling for project portfolio benefit prediction via a GA-BP neural network.
- Source :
- Technological Forecasting & Social Change; Oct2022, Vol. 183, pN.PAG-N.PAG, 1p
- Publication Year :
- 2022
-
Abstract
- Project portfolio benefit (PPB) prediction can effectively help managers monitor the acquisition of PPBs, thereby better achieving their target benefits. However, no valid model is available for PPB prediction. To narrow this research gap, we develop a model based on a backpropagation neural network improved with a genetic algorithm (GA-BPNN) to quantitatively forecast PPBs. First, the evaluation criteria for benefits are determined. Second, the input and output variables of the model are determined and calculated. Third, the initial weights and thresholds of the BPNN are improved by the GA. Fourth, based on the above optimization results, the GA-BPNN model is trained and tested. Last, the numerical example is provided to demonstrate the application of the proposed model. The results indicate that the established model is feasible and effective in predicting PPBs, with an average prediction accuracy rate of 98.64 %, which is 18.24 % better than that of the base BPNN. The model proposed in this paper effectively realizes the quantitative prediction of PPBs, enriching the research on project portfolio management (PPM) and providing managers with a tool to effectively predict PPBs. • Project portfolio benefits are divided into project benefits and synergy benefits. • A hybrid model via GA-BPNN is proposed to predict project portfolio benefits. • The applicability of the proposed model in uncertain environments is verified. • BPNN and GA are combined and first adopted into the project portfolio field [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 00401625
- Volume :
- 183
- Database :
- Supplemental Index
- Journal :
- Technological Forecasting & Social Change
- Publication Type :
- Academic Journal
- Accession number :
- 158818283
- Full Text :
- https://doi.org/10.1016/j.techfore.2022.121939