Lessons Learned in Applying Domain-Independent Planning to High-Speed Manufacturing

Much has been made of the need for academic planning research to orient towards real-world applications. In this paper, we relate our experience in adapting domain-independent planning techniques to a real industrial problem. We present a simpler formulation of a temporal planning graph-style heuristic, show how to extend it to take resources into account, and evaluate its importance in practice. We also derive several general lessons from our experience which might guide researchers looking to increase the relevance of their work or industrial practitioners seeking to apply planning research to real problems.