A Nongradient and Parallel Algorithm for Unconstrained Minimization
Daniel Chazan, W. L. Miranker · SIAM Journal on Control · 1970
The purpose of this paper is to describe an algorithm for unconstrained optimization which is suitable for execution on a parallel computer. A nongradient method similar in nature to Powell’s method is used and it is shown that the algorithm terminates at the minimum for quadratics and converges for strictly convex twice continuously differentiable functions.