Abstract: A state-event engine manages behavior over time in response to events in deterministic fashion from user-supplied rules and policies for a set of dynamic objects. The campaign engine selectively generates and stores a campaign population, representing different types of marketing campaigns. A collection of dynamic object data stores maintains data on the corresponding marketing agent population, resource population, and listing population. A matching process orders members of the listing populations based on the target of at least two members of the campaign population such that a set of offers to buy and offers to sell the same resources is created. A prediction engines processes historical data to predict how campaigns can best match buyer to seller. Software agents negotiate on behalf of buyer and seller identify potential deals. The system is capable of being implemented using computer network and web-based technology.