A novel approach is proposed for the superstructure-free synthesis and optimization of distributed energy supply systems (DESS) by exploiting the nature of evolutionary algorithms. Current approaches require the designer to define a superstructure, which subsequently is optimized. In the presented method, the a priori specification of a superstructure is avoided: A mutation operator employs generic replacement rules to replace parts of energy supply systems by alternative designs. To minimize both the number of replacement rules and meaningless design alternatives generated during mutation, all energy conversion technologies are classified into the so-called energy conversion hierarchy (ECH). The ECH allows for an efficient definition of all reasonable connections between the regarded components, and the definition of generic replacement rules. Thereby, the hierarchy-supported approach balances richness of the available design space and computational efficiency. In addition, the convenient description of the design space allows for an easy addition of technologies into the optimization problem. The proposed approach thus provides an expandable framework for optimizing DESS. The method is applied to the synthesis of a heating and cooling system. The presented approach efficiently solves retrofit and grassroots design problems. It automatically identifies complex solutions such as trigeneration demonstrating the power of the suggested optimization framework. (C) 2012 Elsevier Ltd. All rights reserved.