基于投影算子的回归神经网络模型及其在最优化问题中的应用

Recurrent Neural Network Model Based on Projective Operator and Its Application to Optimization Problems

研究了一种基于投影算子的神经网络模型.与以前研究投影算子的值域一般是n维欧氏空间中的紧凸子集不同,而是n维欧氏空间中未必有界的闭凸子集,同时目标函数也是一般的连续可微函数,未必为凸函数.证明了所研究的神经网络模型具有整体解轨道,以及当目标函数满足某些条件时解轨道的整体收敛性.此外,还将所研究的模型应用于闭凸约束极小化问题以及非线性互补问题和隐互补问题中,并通过数值模拟说明了该神经网络方法的有效性.

The recurrent neural network（RNN） model based on projective operator is studied. Different from the former study, the value region of projective operator in the neural network which they study was a general dosed convex subset of n demensional Euclidean space and it wasn＇ t a compact convex set in general, that is, the value region of projective operator was probably unbounded. They prove that the network has a global solution and its solution trajectory converges to some equilibrium set whenever objective function satisfies some conditions. After that, the model was applied to continuously differentiable optimization and nonlinear or implicit complementarity problems. In addition, simulation experiments confirm the efficiency of the RNN.