Agent-based simulation can be viewed as methodical advancement and generalization of microscopic modelling styles in object-oriented and discrete event simulation well suited for logistic systems. Existing object-oriented simulation frameworks form an adequate basis for agent-based modelling where general simulation functionality can be reused. Extensions are required in view of the main aspects in agent-based simulation related to spatial and behaviour modelling. This paper outlines the advancement of a discrete event simulation framework for applications of agent-based simulation mainly in logistics. Agent-based simulation is typically applied in microscopic modelling of systems where common actions of autonomously deciding actors, mostly humans are represented. The extension includes a graph-based spatial modeling as well as flexible behavior descriptions in different modes (e.g. UML state charts). As an application of agent-based simulation and the related software tool in the logistics domain a simulation model for sustainable logistic concepts in courier services is presented.