One of the challenges during embedded system design is the application driven design. Due to the application driven design, the objectives that are steering the design of an embedded system are mainly based on the needs of the application(s). Examples of embedded system objectives are performance, power, but also battery lifetime and security. Such, potentially conflicting, objectives severely complicate embedded system design. The system level design methodology reduces the complexity of embedded system design by providing a structured approach to reduce the implementation efforts.Part of the system level design is the early design space exploration. At a high level of abstraction, the design space of potential designs is partially explored to make early design decisions that reduce the effort that is spent in later design phases. In our case, a design is defined by a mapping of a multi-application workload onto an MPSoC architecture. This mapping is evaluated using our high-level simulation framework Sesame that is capable to determine the quality of a single mapping in the order of seconds. There are many different ways of mapping a multi-application workload onto the architecture. Although early design space exploration is already complex, it is becomes even more complex due to the growing dynamism in the embedded systems. In this thesis, two sources of dynamism are investigated: at the application side and at the architecture side. The main goal of this thesis is to extend the design space exploration to take into account the dynamism in the embedded system.