IEEE 12th Signal Processing and Communications Applications Conference, Kusadasi, Turkey, 28 - 30 April 2004, pp.304-307
In this study, based on classical gradient projection method of optimisation theory a dynamic solver for linearly constrained linear optimisation problems called gradient projection network is introduced. To illustrate the performance of the network the result obtained for a special linear problem is compared with those obtained by Kennedy-Chua, Rodriguez-Vazquez and Zak's network.