شماره ركورد كنفرانس :
453
عنوان مقاله :
A modified class of conjugate gradient methods with a new nonmonotone Armijo-type line search
پديدآورندگان :
Nosratipour H نويسنده , Amini K نويسنده Inland Waters Aquatics Stocks Research Center,Gorgan, Iran
كليدواژه :
sufficient descent , global convergence , Unconstrained optimization , Nonmonotone Line Search , Strongly convex function
عنوان كنفرانس :
چهارمين كنفرانس بين المللي انجمن ايران تحقيق در عمليات
چكيده فارسي :
Shi and Guo [1] developed a conjugate gradient method with new nonmonotone line search (SG
method), their line search is very complex which has two inequalities, and its second inequality is used to
generate sufficient descent direction. Here, we modify the SG method by leaving out second inequality. We
show that the search direction generated by the modified method at each iteration satisfying the sufficient
descent condition. We prove the global convergence of the new method
شماره مدرك كنفرانس :
1891451
