Experiment #1 Holding DIM=2 IPFS=10 DOMAIN_SIZE 1000 Unif/Grid size=50, thus 400 piece - held constant Vary Pieces 1000 to 10000 Experiment #2 Holding DIM=4 IPFS=10 DOMAIN_SIZE=1000 Unif/Grid size=250, thus 256 pieces - held constant Vary Pieces 1000 to 10000 //============================================================ // Experiments 3-5 differ only in the percentage of pieces used // in making the initial uniform grid. In Exp#3 very // few are used and grid bound values are not terribly // predictive. Thus V3 outperforms V4 for higher IPFs //============================================================ Experiment #3 Holding DIM=2 PCS=3500 DOMAIN_SIZE=1000 Unif/Grid size 100 - so 100 uniform pieces Vary IPFS 5, 10, ... 75 *Note: too few uniform, ineffective grid bound *For higher IPFS, V3 outperforms V4 Experiment #4 Holding DIM=2 Pieces=3500 DOMAIN_SIZE=1000x1000 Unif/Grid size 50 - so 400 uniform pieces Vary IPS 5, 10, ... 75 Experiment #5 Holding DIM=2 Pieces=3500 DOMAIN_SIZE=1000x1000 Unif/Grid size 20 - so 2500 uniform pieces Vary IPS 5, 10, ... 75 //============================================================ // Experiments 6-7 testing the effect of solution threshold //============================================================ Experiment #6 Holding DIM=2 IPFS=75 Pieces=7500 DOMAIN_SIZE=1000x1000 Unif/Grid size 40 - so 625 unif pieces, 6875 smart Vary Thresh 90, 91, ..., 90, 90.2, ... 90.8, 100 Experiment #6 Holding DIM=4 IPFS=75 Pieces=7500 DOMAIN_SIZE=1000x1000 Unif/Grid size 200 - so 5x5x5x5=625 unif pieces, 6875 smart Vary Thresh 90, 91, ..., 90, 90.2, ... 90.8, 100