Computational Results for an Efficient Implementation of the GOP Algorithm and Its Variants

摘要: Recently, Floudas and Visweswaran (1990, 1993) proposed a global optimization algorithm (GOP) for the solution of large class nonconvex problems through series primal relaxed dual subproblems that provide upper lower bounds on solution. (1995a) reformulation in framework branch bound approach allows an easier implementation. They also implicit enumeration all nodes resulting tree using mixed integer linear (MILP) formulation, branching scheme reduces number from exponential to linear. In this paper, complete implementation new versions GOP algorithm, as well detailed computational results applying various classes is presented. The considered including pooling blending problems, with separation heat exchanger networks, robust stability analysis real parameter uncertainty, concave indefinite quadratic medium size.