Role-Based Access Control (RBAC) remains one of the most popular authorization control mechanisms. Workflow is a business flow composed of several related tasks. These tasks are interrelated and context-dependent during their execution. Under many circumstances execution context introduces uncertainty in authorization decisions for tasks. This paper investigates the role-based authorization model with the runtime context constraints and dynamic cardinality constraints. The Generalized Stochastic Petri-net is used to model the authorization process. Moreover, due to the state explosion problem in the Petri-net formalism, the proposed modeling method combines the Queuing theory to analyze both system-oriented and user-oriented performance. Given the workflow information, its running context and the authorization policies, this work can be used to predict the performance of these workflows running in the system. The prediction information can give insight in how to adjust authorization policies to strike a better balance between security and performance.