Several formalisms and techniques have been introduced in the literature for the purpose of modeling and evaluation of complex systems. Each of them has its strengths and weaknesses, which also depend on the purpose of the evaluation. In this paper we propose the integration of two different formalisms in a single framework for the modeling, validation, and optimization of production-supply problems. In particular, the proposed framework combines Process Graphs (P-Graphs) as the modeling formalism, and Stochastic Activity Networks (SAN) for the analysis and optimization. The integration proposed in this paper extends the capabilities of the P-Graph formalism to include performance and dependability metrics in the optimization process, without hampering the modeling convenience of P-Graphs. The proposed approach is applied to a case study of the optimization of a power supply network.