Automatically exported from code.google.com/p/ai-algorithmplatform

AI Algorithm Platform -- a series of AI agorithms, including convex hull, nearest neighbor, pathfinding and concollision detection. (Written in C#)

We both develop some classic algorithms, and some new algorithms based on our M2M Model, which is an new approach to implement these algorithms efficiently.

Macro-to-micro (M2M) model is an implementation model that inherits the GrC idea and extends it to some additional highly desirable characteristics. In this paper we introduce an effective pathfinding algorithm based on the M2M model. This algorithm takes O(n) time to preprocess, constructing the M2M data structure. Such hierarchical structure occupies O(n) bit memory space and can be updated in O(1) expected time to handle changes. Although the resulting path is not always the shortest one, it can make a trade-off between accuracy and time cost by adjusting a parameter - range value to satisfy various applications. At last, we will discuss the advantages of the M2M pathfinding algorithm (M2M-PF) and demonstrate the academic and applied prospect of M2M model.

Different levels of abstract for the map:

[1] H. Wan, Y. Zhang, S. Luo, R. Liu, and W. Ye, “The M2M Pathfinding Algorithm Based on the Idea of Granular Computing,” in 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology, 2009, vol. 2, pp. 533–540. pdf

[2] H. Wan, Z. Zhang, and R. Liu, “A Parallel Dynamic Convex Hull Algorithm Based on the Macro to Micro Model,” in 2009 2nd International Congress on Image and Signal Processing, 2009, pp. 1–5. pdf