Study On The Optimization Of Construction Time Considering The Dependence Of Durations

In the practice of construction implementation, each work is performed in strict accordance with the schedule, with the absence of uncertainty, will the construction be completed on time delivery of production or put into use. However, there are many uncertain factors in actual construction such as the reworks caused by the quality of the construction which is not in accordance with the standard implementation, damage or repair of construction machinery, the slowdown caused by construction materials not timely supply and force majeure (weather mutations, earthquakes, etc.) and other factors, all of which will seriously affect the construction schedule and result in the delay of schedule, at the same time it will inevitably cause serious economic losses to the construction unit. Dependence between the durations(time of duration) is very common, from the perspective of the dependence to optimize the network schedule which is formulated ahead and to make a more reasonable use of the free time difference of the works, in order to effectively manage the construction schedule and to ensure the construction time.This dissertation, based on the existing time cost optimization theory, measures the ubiquitous dependence between the durations and uses it to update the duration data, then sets up the mathematical model of the time optimization with the constraint condition of cost, at the same time modifies the optimization model with taking the dependence of the works into account, finally uses the genetic algorithm to complete the time optimization. The main contents include:Firstly, basic methods of construction planning the schedule of CPM/PERT technique have been reviewed. Then the Monte Carlo simulation principle is introduced, and the duration distribution of the Beta, normal and triangular distribution is assumed, according to the simulation results, the average time and standard deviation of the three cases are compared, and the final time and the probability of completion are predicted. On the basis of this, the principle of time optimization is drawn out, that is in the case of adequate resources getting cost as the constraint condition calculates the optimized time while corresponding cost is lowest.Secondly, the time optimization question and steps of dependence between the durations is put forward. Specifically first dependence between the durations is measured, then the serial duration dependence is assumed as linear and Pearson product-moment correlation coefficient is adopted to measure it; the parallel duration dependence is assumed as nonlinear, from the perspective of generating the parallel dependence and considering the common influencing factors and the organization management, nonlinear structural equation model is adopted to measure it. Thereafter, the durations in the schedule established by the preliminary are updated, besides the time and completion probabilities are simulated by Monte Carlo.Thirdly, based on the above-mentioned time optimization model, cost is not divided into direct cost and indirect cost but as a whole, the relationship of cost and time is assumed to be quadratic function, and the result of duration dependence measured is thought as the time optimization coefficient to establish the dependent time optimization model. Based on genetic algorithm, the design of chromosomes, the initial population generation, fitness evaluation, genetic operator design and termination conditions are described in detail.Finally, the procedure of the time optimization is demonstrated through a construction example, the practical applicability of the model is checked. By comparing and analyzing the results of the classic time optimization and the time dependent duration optimization based on genetic algorithm, it is found that the time dependent duration optimization is considered to be more effective.