This paper deals with a closed-loop supply chain network and proposes new approaches in metaheuristics and exact methods as solution methods. Moreover, the downside risk is incorporated into the objective functions as a risk measure. Hence, the developed two-stage stochastic model aims to minimize the expected total cost and the downside risk, simultaneously. Besides, this study presents the closed-loop network which considers the forward and reverse networks in an integrated manner. In the forward network, this case just considers the forward network, while, the reverse logistic fully focused on the backward network by considering recovering centers (i.e. from recovering centers to remanufacture, recycling and disposal centers). In order to address the problem, ICA, PSO, GA, and also ɛ-constraint method, are utilized. In addition, the parameters of algorithms are tuned by Response Surface Method (RSM) with an MODM approach. To explain the efficiency and effectiveness of methods, four assessment metrics are introduced. At the end, the results show the capability of ICA through the most of the tests problem. According to the risk management, a real data set is used to do some sensitivity analyses on the proposed model.