Interactive Strategy-Based Validation of Behavioral Models

Abstract

When behavioral models are derived automatically based on observed stakeholder interactions, requirements engineers need to validate whether the stakeholders agree with the synthesized behavioral models. Allowing stakeholders to experience such models through simulation and animation allows them to comment on, amend to and correct these models. However, to ensure an efficient stakeholder validation, the simulation has to be guided instead of confronting the user with random situations over and over again.

In this paper, we present a strategy-driven simulator capable of guiding the execution of behavioral models based on graph transformations. By analyzing either the overall structure of a partial state space (look ahead) or by performing an in-depth analysis of the states therein, the simulator is able to determine which transformations should be executed next to continue on the most promising path through the overall state space. The discussed implementation is illustrated with a case study.