Implementation of Machine Learning Models in the Construction Site Organization Process

Introduction. Strategic planning is important for effectiveness of investment and construction projects. Complexity of solving problems increases with growth of variables in calculations and constraints. For effective planning of construction sites, many factors need to be taken into account, such as spatial constraints and distances between objects. To solve such problems, it is possible to use a mathematical machine learning model – a genetic algorithm (GA) to optimize the placement of objects on a construction site. The aim of this study is to improve the accuracy and flexibility of solutions in the organization of the construction site, reduce complexity of calculations and minimize the amount of data. Materials and Methods. The realization of this goal is possible by introducing the Systematic layout planning (SLP) method into the planning process to optimize space on construction sites. To confirm the effectiveness of optimizing the location of objects, the SLP method was applied in the organization of planning the construction site of an administrative building. The planning took into account the stages of work, the required economic facilities, data on safety and environmental safety of the construction site. The use of the Dynamo plugin for analysis made it possible to adjust the location of objects taking into account the utilization factor of the territory. Research Results. As a result of the modeling of the construction site plan, it was found that using the SLP method, the process is adapted, taking into account the location of objects according to the established values of the relationship matrix using automation and color coding to simplify analysis. Flexibility in making informed decisions is important for designers, given the safety and intensity of the workflow. The SLP method reduces distances through optimization of logistics, optimizes the location of objects on the construction site, taking into account restrictions. This hybrid approach increases the efficiency of implementing machine learning models in the design process of building master plans of facilities. The integration of the genetic algorithm and BIM technologies into the construction site organization process helps to optimize solutions based on optimal distances, improves the visualization of decisions and problem correction. Discussion and Conclusions. The results of the study contribute to decision-making efficiently and quickly, minimizing the time required for analysis compared to some other approaches. The result of the research is the creation of a system that is not only flexible and adaptable, but also overcomes the limitations typical of previous methods. The use of machine learning technologies to predict optimal design decisions and automate the establishment of key relationships can significantly reduce the need for manual data entry, thereby simplifying and speeding up development processes. Adding new examples of diverse plans and spatial constraints will strengthen the foundations of the concept and enrich its application in a variety of investment and construction projects.