بهینه سازی مشخصات جاده جنگل با استفاده از تکنیک های فراشناختی

An optimal design of vertical alignment, considering the design constraints and costs is one of the most complicated problems of road planning and construction. The results of many linear, nonlinear and heuristic techniques that can enhance design ability to minimize the total cost of road construction using many different variables are well acknowledged. It is assumed that the genetic algorithm (GA) and Particle Swarm Optimization (PSO) can be efficiently applied for road vertical alignment allocation. This paper focuses on solving vertical alignment optimization problem using meta-heuristic algorithms. Two intelligent optimization tools of GA and PSO have been used to find a near optimal forest road profile, connecting specified endpoints considering restrictions associated with forest road profile design with cost evaluation. A number of setting parameters such as population size and crossing over and mutation rate in GA and also best group and particle's position in PSO were tested to search the global optimal answer. Results of optimization by GA and PSO approaches were compared with the common manual road profile drawing method. Results indicated that the GA and PSO could reduce earth work volume costs while designing more smoother and qualified alignment in comparison with the manual design. Results suggested that among the applied optimization methods, the GA was the most suitable one for this feature of the problem since it is able to save optimum position at better solutions with a reduced computed cost. From the cost point of view, it was cleared that optimizing the fixed length of road profile applying GA, with different population size, would be better for big numbers of control points but smoother for low numbers of control points.