The investment to DERs (Distributed Energy Resources) in the emerging smart energy grids are critical to achieving goals for environmental protection, reducing energy costs and coping with failures. An architecture and market mechanism is needed to support the participation of owners of DERs that aim to maximize the profit from this investment within a system that is also able to achieve grid level objectives. Due to privacy requirements, owners of DERs should not be requested to divulge information such as the state of charge of a local energy store, and yet such information is very relevant for operating the grid in such a way that meets the objectives of the owner of the grid, who may operate a large power plant that is ultimately responsible for satisfying the load that cannot be covered by DERs. This paper builds on previous work to propose a modular multi-agent architecture in which prosumer agents represent DER owners and an auctioneer agent represents the power plant operator. The architecture allows the different agents to use different algorithms aiming at satisfying local and grid level objectives while meeting the privacy requirement. In particular, a fuzzy logic based algorithm is developed for the distributed automation platform that supports the architecture. The capability of the architecture to flexibly evaluate the performance of various algorithms from the DER owner and plant operator perspective is demonstrated with a case study in the district heating domain.