美国数学建模竞赛特等奖论文——mcm2003_A_wang_deng_liu【数学建模】.pdfVIP

Team # 167 Page 1 of 11 Shelling Procedure and Optimization by Simulated Annealing For Sphere Packing Summary We provide two models (model A and model B) for Gamma Knife treatment planning, one is a Mixed Integer Programming (MIP) model, and the other is an optimal sphere-packing model. Based on dose distribution and the requirements of Gamma Knife unit, a Mixed Integer Programming (MIP) model is constructed and discussed. Another model is a sphere packing approach to Gamma Knife Treatment Planning problem. Based on the theory of digital image processing and simulated annealing, an algorithm to obtain an initial sphere-packing plan is designed, and a method from the thought of simulated annealing is used to optimize the solution. Due to the complexity of the MIP model and the large variables, the calculating would cost too much time. Many researchers have developed or have been developing other programming models. There are the same problems of time complexity. So we have to establish model B to speed up the computation and find the best or approximate best plan for the Gamma Knife Treatment. In model B, we simulate some images from CT/MRI, and then translate them into digital-image as matrices by image processing. Then we get an initial result through Shelling Procedure and adjust further. Our algorithm is recursion procedure for sphere packing. It determines the candidate location of the sphere center heuristic from largest sphere radius to the smallest one. As we know, when a ball moves inside a large close region, the track of the sphere center forms a close surface with the shape similar to the boundary of the region. We adopt heuristic method to choose one voxel as the ce