The emphasis in this paper is on investigating a new generation scheduling algorithm for the interconnected power systems. In general, the generation scheduling problem formulated as a mixed integer nonlinear programming (MINLP) can be efficiently computed by the generalized Benders decomposition (GBD) technique which decouples an original problem into the master problem and subproblems to allow remarkably fast and accurate solutions of very large problems. In order to ferret out efficient inter-temporal optimal power flow subproblems, we will propose a regional decomposition framework based on auxiliary problem principle (APP). Obviously, this scheme can find the most economic dispatch (ED) schedule under the power transactions for a multi-utility system without the exchange of each utility's own private information and major disruption to existing ED or optimal power flow (OPF) constructed by individual utilities. (C) 2010 Elsevier Ltd. All rights reserved.