An Extensible Framework for Real-Time Task Generation and Simulation

In real-time systems research, validations are usually performed by executing synthetically generated tasks through programmatic implementations of derived algorithms or theoretical results. For every new result, real-time researchers have to develop systems, several times from scratch, for generating task sets as well as implementing their derivations. Another issue arises when the results are submitted for peer review. Reviewers only have access to results in the form of numerical values given in the paper, and have no easy way of validating the results themselves. To solve these two issues, we present a new extensible system for real-time task generation and simulation. Using modern software engineering principles of object and reflection-oriented programming, we show how real-time analysis can be partitioned into sub-systems, where each such subsystem can be implemented as a run-time ´plug-in´ which can be independently developed by different research groups. This technique is intended to save real-time researchers the significant amount of time spent in result validation, as well as allow reviewers easy access to the experimental setup of the submitted paper for a more efficient review process.