Context. Many modern software applications have a cloud-based architecture and can be scaled out and in to match fluctuating user demand. Such architectures can be modeled in the Palladio Framework prior to their deployment. This enables assessing their qualities and improving their configurations if necessary. Problem. Currently, finding the best set of scaling rules requires manually defining and evaluating di!erent configurations and comparing their qualities. This manual process is not only error-prone, but can also lead to suboptimal results. Objective. The objective of this work is the development of a tool for the procedural optimization of the so-called elasticity policies of a given software architecture, i.e. the rules defining if, where and how the architecture should be scaled in order to match the user demand. Method. The objective is achieved by extending PerOpteryx - a tool from the Palladio Framework for the optimization of non-elastic software systems, to enable the optimization of elasticity policies. This is achieved by combining it with the Slingshot Simulator - a tool for modeling and evaluating software system elasticity. Result. The developed tool is validated using two di!erent software architectures. Starting with an initial configuration of each architecture, an improvement to the cost and performance is achieved in both cases by optimizing their respective elasticity policies. Conclusion. Based on the results it is concluded that the developed tool makes the optimization of the elasticity policy of a given software system at its design time possible.