Improved GA for construction progress and cost management in construction projects

Abstract The construction progress and cost management (CPCM) of construction projects are key areas that need to be focused on when formulating construction plans and conducting on-site commands. A CPCM method is proposed based on the standard genetic algorithm. During the process, a serial operation strategy is introduced, a project valuation function is defined, and a CPCM resource library is constructed. The tangent function is used to simulate the key relationships in the construction process, and finally, the construction progress and cost are compressed and optimized. The experimental results indicate that in optimal fitness tests on the construction engineering cost standards (CECS) dataset, the proposed method reaches the upper limit of 20.9 in just 23 generations, 66.67% faster than multi-stage genetic algorithm’s (MGA’s) 72 generations and 51.06% faster than quantum genetic algorithm’s (QGA’s) 47 generations. For the building engineering dataset (BED) dataset, it achieves the upper limit of 21.2 in 30 generations, 33.33% faster than MGA’s 45 generations and 49.18% faster than QGA’s 61 generations. In MSE change tests, the proposed method’s MSE fluctuates within a range of 0.04 after 22 and 23 generations for the CECS and BED datasets, respectively, which is more stable than MGA and QGA. Additionally, when the number of engineering steps increases from 5 to 25, the calculation time of the proposed method increases by less than 30 s, demonstrating higher computational efficiency than that of MGA and QGA. When conducting cost compression, the research method generates a plan that reduced construction costs by 230 k yuan. This indicates that the research method has good computational efficiency and can effectively generate the reference of project construction cost management plans.