Acceleration method for convex optimization over the fixed point set of a nonexpansive mapping (Q2515036)

This paper concerns the following convex optimization problem: given a convex, continuously Fréchet differentiable function \(f\) on a real Hilbert space \(H\) and a nonexpansive (that is, \(1\)-Lipschitz) self-mapping \(N\) of \(H\) with a nonempty fixed point set \(\mathrm{Fix}(N)\), minimize \(f(x)\) subject to \(x \in \mathrm{Fix}(N)\). Such problems arise, for example, in signal recovery, beamforming and bandwidth allocation. The author proposes an algorithm the aim of which is to accelerate both the objective function minimization as well as the search for a fixed point of the nonexpansive mapping. He presents a comprehensive convergence analysis of this algorithm and compares it with several existing ones. Detailed numerical comparisons of the proposed algorithm with existing fixed point optimization algorithms are also provided. These comparisons demonstrate the effectiveness of the proposed algorithm.