Convergence and error bound of a D-gap function based Newton-type algorithm for equilibrium problems
暂无分享,去创建一个
The D-gap function approach has been adopted for solving variational
inequality problems. In this paper, we extend the approach for
solving equilibrium problems. From the theoretical point, we study
the convergence and global error bound of a D-gap function based
Newton method.
A general equilibrium problem is first formulated as an equivalent
unconstrained minimization problem using a new D-gap function. Then
the conditions of "strict monotonicity" and "strong monotonicity"
for equilibrium problems are introduced. Under the strict
monotonicity condition, it is shown that a stationary point of the
unconstrained minimization problem provides a solution to the
original equilibrium problem. Without the assumption of Lipschitz
continuity, we further prove that strong monotonicity condition
guarantees the boundedness of the level sets of the new D-gap
function and derive error bounds on the level sets. Combining the
strict monotonicity and strong monotonicity conditions, we show the
existence and uniqueness of a solution to the equilibrium problem,
and establish the global convergence property of the proposed
algorithm with a global error bound.