磁性目标定位问题可归结为一类非线性规划问题的求解 .该问题的最大特点就是其目标函数的计算过程极为繁琐 .而目标函数计算的快慢对磁定位的实时性有很大影响 .在详细研究了目标函数之后 ,给出了目标函数的一种神经网络结构实现 .由于该网络结构易于并行计算和VLSI实现 ,从而可使磁定位的实时性得到改善 . The problem of magnetic target localization can be attributed to the solution of a class of nonlinear programming problems.The most prominent feature of this problem is that the calculation of its objective function is very complicated and the calculation of target function has a great influence on the real-time performance of magnetic localization. After studying the objective function, a neural network structure of the objective function is given, which can improve the real-time performance of the magnetic localization because of its easy parallel computation and VLSI implementation.