The computational complexity of distance functions of two-dimensional domains

We study the computational complexity of the distance function associated with a polynomial-time computable two-dimensional domains, in the context of the Turing machine-based complexity theory of real functions. It is proved that the distance function is not necessarily computable even if a two-dimensional domain is polynomial-time recognizable. On the other hand, if both the domain and its complement are strongly polynomial-time recognizable, then the distance function is polynomial-time computable if and only if P=NP.