The paper describes software combining parametric geometric modeling with a version of the harmony search method, modified to support multimodal structural optimization. Researchers have recognized the potential of population-based optimization methods, such as genetic algorithms, to support multimodal optimization: that is, generating a diverse range of good alternative solutions, rather than a single best solution. Among these methods is the harmony search method, which has been demonstrated to be efficient in many unimodal structural optimization problems. The paper describes a new version of the harmony search method, implemented as an assembly within Bentley's Generative Components, enabling high-level control of geometry. The new method is demonstrated on an bridge supported by two inclined parabolic arches, a structure where GC controls a complex geometry with a single variable. Comparative studies of the example find that the new method is more effective than conventional harmony search in consistently finding multiple good solutions.