File(s) under permanent embargo
Prediction and improvement of labor productivity using hybrid system dynamics and agent-based modeling approach
journal contributionposted on 2018-01-01, 00:00 authored by M Khanzadi, Farnad NasirzadehFarnad Nasirzadeh, M Mir, P Nojedehi
Purpose: The purpose of this paper is to present a hybrid simulation approach for predicting the value of labor productivity taking account of various continuous influencing factors and the interactions between different agents involvedin the project. Design/methodology/approach: The various continuous factors affecting labor productivity are simulated using system dynamics (SD). The heterogeneity of different agents involved in the project and their interactions is accounted using agent-based modelling (ABM). The developed ABM and SD models are finally integrated to simulate the value of labor productivity taking account of all the influencing factors. Findings: The proposed hybrid simulation tool is implemented in a real project to evaluate its perfomance. The value of labor productivity is simulated by taking account of all the influencing factors. The most appropriate execution strategy is then selected using the developed hybrid SD-ABM approach to improve productivity. It is shown that the number of working groups and their movement patterns affect the severity of the groups' interferences which will in turn affect the value of labor productivity. Practical implications: This research helps project managers to predict and improve the value of labor productivity taking account of all the influencing factors. Originality/value: It is believed that the proposed hybrid SD-ABM simulation approach offers a novel and robust tool for modeling labor productivity because the effects of various continuous influencing factors and the interactions between different agents are taken into account through the combination of SD and ABM. Many complex problems faced in construction projects involve interacting elements of a different nature, and the integration of SD with ideas from ABM offers potential to combine the strengths of the two methodologies to solve the problem.