Two-dimensional data sequence and its grey generation

Comparing to the general sequence and its generations in grey system theory (GST), this paper defines the two-dimensional data sequence, shows the weakness of average generation of one-dimensional data sequence and proposes an idea of using concave and convex properties of discrete data in local sequence to build up the criteria of choosing weight coefficient for the preference generation. For two-dimensional data sequences, the weight coefficient should be decomposed into corresponding axis for obtaining a real preference generation. Since the elements in two-dimensional data sequence may have their different expressions, the same accumulating generation operation would cause different results generated and some would have divergent meanings. This paper discusses the accumulating generation of two-dimensional data sequence and defines a new kind of accumulating operator. The approach proposed in this paper can also be extended to the case of three-dimensional data sequence.