Research on construction engineering layout algorithm based on action space optimization

Traditional layout optimization methods are mostly based on two-dimensional rectangular packing problem, but many factors such as structural requirements, safety specifications and construction convenience need to be considered in practical engineering, so more advanced optimization algorithms are needed. In this paper, an architectural engineering layout algorithm based on action space optimization is proposed, aiming at exploring a more efficient layout scheme by optimizing action space. Firstly, the building space is divided into grids, the action space set is initialized, and the action space is dynamically updated according to the arranged rectangular blocks. Combining the greedy layout strategy in the lower left corner and the backtracking optimization method, the local optimal solution problem is effectively avoided by comprehensively considering the space utilization rate and the search depth. The algorithm adopts 3D tensor coding action space, combines the mixed strategy of best-first search and depth-first search, controls the complexity through pruning and memory storage technology, and introduces the constraint of safe distance and optimization of construction streamline to enhance the engineering adaptability. In the experimental verification part, three groups of typical architectural engineering layout examples are selected, and the pure greedy algorithm without backtracking and the algorithm in this paper are compared. The results show that the algorithm in this paper significantly improves the space utilization rate, especially in complex cases, by more than 20%. At the same time, by dynamically updating the action space, the compliance rate of safety spacing constraints is 100%, and the material error rate is controlled within 3%. The optimization effect of construction streamline is remarkable, and the average transportation distance and the number of intersection points of paths are greatly reduced. Robustness test shows that the algorithm has good stability and reliability under different initial conditions.