Within a medium-term market horizon, this research work provides a methodology that allows a commercial virtual power plant (CVPP) to form an optimal coalition of heterogeneous distributed energy resources (DERs) based on weekly bilateral contracting, futures-market involvement, and pool participation. The established model aims at composing an optimal portfolio of available DERs and jointly takes into account the risk associated with the energy trading strategy of the CVPP. Perceiving the fact that pool prices have highly uncertain nature, a framework based on stochastic programming approach is utilized to model this decision-making problem. The proposed framework consists of two stages. The first stage deals with decisions regarding DERs optimal selection for the VPP coalition, the amount of agreed quantity in the bilateral negotiation, and the type and quantity selection of futures-market contracts as well. In the second stage, decisions are made based on the most plausible realizations of the stochastic prices in the day-ahead market. For a given pre-specified risk level on profit volatility, the main objective is to maximize the expected profit for the VPP manager over the planning horizon. The efficiency and applicability of the developed model is illustrated and analyzed by its implementation in a system with few heterogeneous DERs and through different scenarios, and finally thereby meaningful conclusions are duly drawn. (C) 2016 Elsevier Ltd. All rights reserved.