A close encounter with Random Numbers

Abstract

Performance profiling or empirical testing, and statistical testing of algorithms for NP-complete problems is typically based on random sample testing. Random values constitute a good source of data for testing the effectiveness of a computer algorithm. Random number generation is an absolute proposition. As such, generally the concentration is on realistic pseudorandom number generation. There are numerous pseudorandom number generation algorithms. We propose here another drop in the sea which is at least as efficient as the existing algorithms and simpler in certain respects.