The dynamic economic environment is driving the evolution of traditional supply chains toward a connected, smart, and highly efficient supply chain ecosystem. Algorithms become powerfull tools that enable machines to make autonomous decisions in the digitized supply chain of the future. The integration of software agents with decision support systems provides automated means for decision making. The present paper proposes an agent-based decision support system for supply chain management in the petroleum industry. This industry has a strategic position as it is the base for other essential activities of the economy of any country. The petroleum industry is faced with volatile feedstock costs, cyclical product prices and seasonal final products demand. The current paper considers the position of a refinery as it is at the middle of the integrated petroleum supply chain, between the upstream and downstream. It procures crude oil from upstream assessing the price, quality, timing, and distance to the refinery in order to decide the optimal acquisition. Additionally, the refiner has to carefully monitor the price risk and manage the inventory. The manufacturing activities of the refiner requires thoroughly planning and scheduling the production levels and supply chains for all the derivates and feedstocks for petrochemical industry using tools for decision making in order to estimate market opportunities and threats under volatile market conditions. In order to provide a reliable and practical decision making model, we proposed the integration of supply chain formation algorithm and a mechanism for decision support under uncertainty using maximum expected utility.