Stochastic Optimization of Construction Projects Planning with Genetic Algorithm

This paper presents a stochastic approach for optimization of significant parameters of project control such as duration and budget in construction projects with Genetic Algorithm (GA). Increasing of uncertainty in project executions encourages the project planners to model construction projects by stochastic optimization methods. In this study, GA is considered as a controlled Markovian process in selection, mutation, and crossover stages which are incorporated into an explicitly given transition matrix. The non-stationary Markovian chain representation of GA is applied to construction project as an example to optimize total project cost and duration with stochastic critical path method. We also consider the asymptotic of the steady state distributions as population size increases. The proposed approach improves the project planner estimation during the project feasibility study stage to avoid project delays and over budget.