Increasing the Numeric Expressiveness of the Planning Domain Definition Language

The technology of artificial intelligence (AI) planning is being adopted across many different disciplines. This has resulted in the wider use of the Planning Domain Definition Language (PDDL), where it is being used to model planning problems of different natures. One such area where AI planning is particularly attractive is engineering, where the optimisation problems are mathematically rich. The example used throughout this paper is the optimisation (minimisation) of machine tool measurement uncertainty. This planning problem highlights the limits of PDDL's numerical expressiveness in the absence of the square root function. A workaround method using the Babylonian algorithm is then evaluated before the extension of PDDL to include more mathematics functions is discussed.