The art of computer programming, volume 1 (3rd ed.): fundamental algorithms
The art of computer programming, volume 1 (3rd ed.): fundamental algorithms
| Hi-index | 0.00 |
A monte-carlo simulation was utilized to determine the best values for the reallocation coefficients that are associated with the modified real-location algorithms1 whose original versions are given by Garwick and Knuth.2 Various distributions of region sizes were dynamically created during the insertion-deletion process, and the optimum real-location coefficients were found. Then the distributions were changed in the midst of the insertion-deletion process, and a different set of optimum coefficients were discovered. Global overflow was assumed to occur at 3 to 5 percent memory remaining, which is optimum.1 To smooth out the simulation, 30 runs were made at each global overflow point, with overflow points occurring at .5 percent intervals. Ten regions were used, and l000 total words of storage were allocated to all of the regions. The simulation was run on the 360/65 at the University of South Carolina, and required 15 hours of CPU time.