Genetic Programming Instructions for Simple Hyperspace Partitioning

M. Šprogar and D. Verber (Slovenia)


Genetic programming, embedded systems, fault detection, monitoring cell.


The article discusses a genetic programming setup in cre ating simple fault spaces by divisioning the search space into a set of exclusive hypercubes, which can be used to recognize faults during system operation. Choice of differ ent functions and terminals greatly affects the convergence speed of the algorithm. The fault-detection problem is ap proched by the use of monitoring cells, which implement the system supervising logic. Genetic programming was chosen because it is the only paradigm with the potential to discover new types of fault-detection rules, what is of great importance for the future research in fault-detection.

Important Links:

Go Back