神经网络和遗传算法在相关峰判读中的应用

在相干光学目标识别技术研究中,为了更好地判读相关峰,把遗传算法和人工神经网络相结合,以反向传播神经网络和遗传算法为基础,建立了以遗传算法优化神经网络初始权值和阈值的相关峰判读系统,既避免了神经网络训练中容易陷于局部最小值和收敛速度较慢的缺点,又克服了遗传算法局部精确搜索能力的不足,实现其优势互补,从而有利于更好地解决相关峰判读问题。结果表明,改进后的算法充分发挥遗传算法和反向传播算法的优点,达到了较好的判读效果。

Application study on neural network and genetic algorithm in the interpretation of correlation peak

In order to identify correlation peak much better in the research of targets recognition technology by coherent optics, the method of combining genetic algorithm (GA) with artificial neural network (ANN) was introduced. Based on this, the correlation peak identification system was built by using GA to optimize the initial weights and thresholds of ANN, witch could avoid the defects both from ANN of trapping into local minimum and low convergence speed and from GA of bad local searching precision. It was propitious to solve the problem of recognizing correlation peak. The testing results showed that the improved method, which got much better recognition effects, made good of the advantages of GA and ANN.