In decision support systems, it is essential to get a candidate solution fast, even if it means resorting to an approximation. This constraint introduces a scalability requirement with regard to the kind of heuristics which can be used in such systems. As execution time is bounded, these algorithms need to give better results and scale up with additional computing resources instead of additional time. In this paper, we show how multi-agent systems can fulfil these requirements. We recall as an example the concept of Evolutionary Multi-Agent Systems, which combine evolutionary and agent computing paradigms. We describe several possible implementations and present experimental results demonstrating how additional resources improve the efficiency of such systems.