2-D lattice CA and configurable CA for pseudorandom number generation

Abstract

<P>IN THIS THESIS, A NEW 2-D CA STRUCTURE IS PROPOSED A?? THE 2-D LATTICE CA, WHICH AIMS TO DESIGN A GENERALIZED FORM OF 2-D CA GRID PRNG THAT OFFERS BETTER PERFORMANCE. A 2-D CA LATTICE CA STRUCTURE WHICH IS SIMPLER AND MORE COST-EFFECTIVE THAN THAT OF THE 2-D CA GRID CAN GIVE GOOD RANDOMNESS QUALITY, BETTER THAN THE PERFORMANCE OF 2-D CA GRID PRNG AND 1-D CA PRNG. A NEW METRIC #RN IS INTRODUCED AS A WAY OF FINDING FEWER CELLS IN CA THAT CAN PASS THE RANDOMNESS TESTS. RELATED TO THE METRIC #RN, A CROPPING TECHNIQUE IS ALSO PROPOSED TO FURTHER IMPROVE THE RANDOMNESS QUALITY OF 2-D LATTICE CA PRNGS AND OTHER CA PRNGS. FINALLY, CONFIGURABLE CA (COCA) IS PRESENTED TO PROVIDE A GENERAL IDEA OF CA, WHICH CAN BE APPLIED NOT ONLY IN RANDOM NUMBER GENERATION BUT IN MODELING SYSTEMS AS WELL. </P>