-
Notifications
You must be signed in to change notification settings - Fork 0
/
experiments_emo21_2d_final.R
142 lines (112 loc) · 4.58 KB
/
experiments_emo21_2d_final.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
## Code for 2d experiments
library(batchtools)
library(checkmate)
################################################################################
## calculate improved latin hypercube designs for the experiments
# set.seed(1)
# ns = c(50, 100, 200, 500, 1000, 2000, 5000, 20000)
# ilhds = vector("list", 8)
# for(i in 1:8) {
# ilhds[[i]] = generate_lhd(n = ns[i], lower = c(-pi, -pi), upper = c(pi, pi),
# type = "improved")
# }
# names(ilhds) = ns
# save(ilhds, file = "ilhds.RData")
################################################################################
set.seed(1)
reg = makeExperimentRegistry(file.dir = "experiments_emo21_2d_final")
generate_data = function(data, job, n) {
source("functions_experiments_2d.R")
library(checkmate)
## generate design:
load(file = "ilhds.RData")
xv2d = ilhds[[which(names(ilhds) == n)]]
cv2d = rmvnorm_par(n = n, par = list(mean = c(50, 70),
sigma = matrix(c(9, -6, -6, 9), nrow = 2)))
design2d = design_xc(x = xv2d, c = cv2d)
## evaluate design:
data2d = evaluate_design(design = design2d, f = f_artificial_2dx2dc)
## build model:
model2d = build_kriging_model(data = data2d)
## convert model to f:
f2d = model_to_f(model = model2d)
## x-values for evaluation:
xv2d = ilhds[[8]]
return(list(xv2d = xv2d, cv2d = cv2d, f2d = f2d))
}
integration_2d_known = function(job, data, instance, level = 10, ncpus = 10) {
source("functions_experiments_2d.R")
library(checkmate)
library(mvQuad)
library(parallelMap)
## estimate parameters:
mean2d = colMeans(instance$cv2d)
sigma2d = cov(instance$cv2d)
## split x-values for different cores:
nx = nrow(instance$xv2d)
sx = split(instance$xv2d, rep(1:ncpus, each = ceiling(nx / ncpus))[1:nx])
## integration limits:
il2dk = int_lim(values = instance$cv2d)
## export:
parallelLibrary("checkmate", "mvQuad")
parallelExport("instance", "level", "sx", "mean2d", "sigma2d", "il2dk", "test_dmv",
"dmvnorm_par", "rmvnorm_par", "design_xc", "evaluate_design",
"f_artificial_2dx2dc", "build_kriging_model", "model_to_f")
## parallel mvQuad:
i2dk = parallelLapply(sx, function(y) {
r = apply(y, 1, function(z) {
tmp = mvQuad_ev(f = instance$f2d, a = il2dk$lower, b = il2dk$upper,
x = matrix(z, nrow = 1), pdf = dmvnorm_par,
par = list(mean = mean2d, sigma = sigma2d), level = level,
type = "cNC4")
return(tmp)
})
do.call(rbind, r)
})
i2dk = do.call(rbind, i2dk)
return(i2dk)
}
integration_2d_unknown = function(job, data, instance, level = 10, ncpus = 10) {
source("functions_experiments_2d.R")
library(checkmate)
library(mvQuad)
library(parallelMap)
## estimated density:
edk2d = edk(values = instance$cv2d)
## split x-values for different cores:
nx = nrow(instance$xv2d)
sx = split(instance$xv2d, rep(1:ncpus, each = ceiling(nx / ncpus))[1:nx])
## export:
parallelLibrary("checkmate", "mvQuad")
parallelExport("instance", "level", "sx", "edk2d", "rmvnorm_par", "design_xc",
"test_dmv", "evaluate_design", "f_artificial_2dx2dc",
"build_kriging_model", "model_to_f")
## parallel mvQuad:
i2du = parallelLapply(sx, function(y) {
r = apply(y, 1, function(z) {
tmp = mvQuad_ev(f = instance$f2d, a = edk2d$par$bounds$lower,
b = edk2d$par$bounds$upper, x = matrix(z, nrow = 1),
pdf = edk2d$pdf, par = edk2d$par, level = level,
type = "cNC4")
return(tmp)
})
do.call(rbind, r)
})
i2du = do.call(rbind, i2du)
return(i2du)
}
addProblem(name = "generate_data", fun = generate_data, seed = 1)
addAlgorithm(name = "integration_2d_known", fun = integration_2d_known)
addAlgorithm(name = "integration_2d_unknown", fun = integration_2d_unknown)
pdes2d = data.frame(n = c(50, 100, 200, 500, 1000, 2000, 5000))
addExperiments(prob.designs = list(generate_data = pdes2d),
algo.designs = list(integration_2d_known = data.frame()), repls = 100)
addExperiments(prob.designs = list(generate_data = pdes2d),
algo.designs = list(integration_2d_unknown = data.frame()), repls = 100)
## to submit the Jobs to SLURM-Cluster:
# submitJobs(findNotSubmitted(), reso = list(memory = 6000, walltime = 48 * 3600, ncpus = 10, pm.backend = "multicore")))
## to reduce the results after all jobs are finished:
# ujt = unwrap(getJobTable())
# save(ujt, file = "ujt2d_emo21_final.RData")
# res = reduceResultsDataTable()
# save(res, file = "res2d_emo21_final.RData")