By modifying the existing HS conjugate gradient method,an improved HS conjugate gradient method with descent property has been proposed,and the descent property of this algorithm has been demonstrated.Under the stan-dard Wolfe line search conditions,it has been proven that the improved HS algorithm possesses global convergence.Final-ly,through the comparison of numerical experimental results,it was found that the new algorithm enjoys excellent numeri-cal performance.
unconstrained optimizationconjugate gradient methodstandard Wolfe line searchglobal convergence
NETL
NSTL
万方数据
