Abstract: While there are different research groups in the mobile computing community, most research requires mobile action data in terms of user calling and mobility patterns. Since collection of such real data is currently difficult and since most activity data is proprietary [12], researchers need to model this mobile activity data to evaluate their work. In [9] we model parameterized mobile actions in a wireless personal communication service (PCS) network based on three components: human behavioral daily movements cycle (traffic and labor), calling patterns, and topological localities. We have developed a mobile action generator based on the model. Our goal in designing the mobile action generator is to provide a common "benchmark" through which researchers can directly exchange performance results and avoid inaccuracies. We describe our model and the structure of our mobile action generator and present simulation results showing the performance of several call setup protocols.