A Hybrid Optimization Algorithm Based On “GA-A~* ” For Cable Harness Design Of Mechanical And Electrical Products

In the present research on routing problems of cable and piping, the branch structure optimization and the path planning are usually separated, so it’s difficult to get a global optimal solution. A hybrid optimization algorithm is proposed based on the improved A~* algorithm and the improved genetic algorithm. The results of branches path planning are introduced into the evaluation function of the structure optimization of the cable harness, which make branch path planning and the cable harness structure optimization have a concurrent evolution in genetic algorithm, so that we can get a global optimization of structure and path in cable harness design.The main work of this paper are as follows:(1) The cable routing optimization problem is analyzed, requirements of the routing space for A~* algorithm and genetic algorithm are compared. A three-dimensional discretization method based on grid points is proposed. Through the array of small ball, interference detection, the coordinates of obstacle points in the discrete space model is obtained. Finally, the discrete space model of the space to be wired is stored in an array format, and the discrete space model which can be recognized by the algorithm is obtained.(2) In view of the process requirements of cable path planning, the traditional A~* algorithm is extended in two aspects, which are increasing geometrical dimension and routing technology.By increasing the Z coordinates, the A~* algorithm can achieve the path search in the 3D space. By the method of interference detection of the different process structure with the array of spheres,the coordinates of the different processing points are obtained. Finally, the discrete space geometry model is stored in a three-dimensional array, and the discrete space model which can be identified by the algorithm is obtained.(3) A hybrid optimization algorithm based on A~* and genetic algorithm is studied. In view of the local optimization is not the global optimum, which leads to the branch structure design and path planning are separated. The improved A~* algorithm and genetic algorithm are combined to achieve the global optimization of the path and the branch structure. Then, because of the high mortality rate, low convergence rate of the genetic algorithm, the selection, hybridization and mutation operation is improved, and by retain the excellent individuals, the performance of the algorithm is improved. The effectiveness of the proposed algorithm is verified by examples.