快速模拟退火法在分形数据压缩中的应用

Applications of Fast Simulated Annealing in Fractal Data Compression

作为一种非凸性优化技术的快速模拟退火法,可以选出局部极小区域而搜索全局状态,而且比经典的模拟退火法有更快的退火率。为了扩大历史数据的存储以及加快状态监测数据和诊断结果的传输,应用分形插值函数对实验模拟汽轮发电机组故障数据进行了压缩。采用快速模拟退火算法来计算分形插值函数的关键参数——纵向比例因子,进一步加快了数据压缩的速度。

As a nonconvex optimization technique, fast simulated annealing may escape from local minimum valleys, and searches for the ground state. Meanwhile, the cooling schedule of fast simulated annealing is fast compared with the classical simulated annealing. In order to enlarge the storage capacity for historic data and to quicken the transimission of condition monitoring data and diagnostic results, fractal interpolation function is used to compress experimental data of turbo-generator simulated faults. Fast simulated anneal-ing is used in calculating the key parameter of fractal interpolation function - vertical scaling factor, thus makes further improvement of data compression speed.