-
Notifications
You must be signed in to change notification settings - Fork 2
/
run_simulator2.m
151 lines (114 loc) · 3.75 KB
/
run_simulator2.m
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
143
144
145
146
147
148
149
150
clear;
lambda = [13 13 13 50 50 50];
S = [1 1 1 3 3 3];
W = [0 60 80 0 180 240];
%ex3_a(lambda, S, W)
ex3_b()
%ex3_c()
function ex3_a(lambda, S, W)
p = 0.4;
invmiu = 90;
Ms = 2;
Mh = 5;
R = 10000;
N = 1000;
runs = 40;
test_count = size(lambda,2);
b_s = zeros(test_count, runs);
b_h = zeros(test_count, runs);
b_s_confidence = zeros(test_count,2);
b_h_confidence = zeros(test_count,2);
for test_nr=1:size(lambda,2)
for lap=1:runs
[b_s(test_nr,lap), b_h(test_nr,lap)] = simulator2(lambda(test_nr), p, invmiu, S(test_nr), W(test_nr), Ms, Mh, R, N);
end
[b_s_confidence(test_nr,1), b_s_confidence(test_nr,2)] = confidence_level(0.1, b_s(test_nr,:), runs);
[b_h_confidence(test_nr,1), b_h_confidence(test_nr,2)] = confidence_level(0.1, b_h(test_nr,:), runs);
fprintf('%.6f +- %.6f || %.6f +- %.6f\n', b_s_confidence(test_nr,1)*100, b_s_confidence(test_nr,2)*100, b_h_confidence(test_nr,1)*100, b_h_confidence(test_nr,2)*100);
end
end
function ex3_b()
S = 2; % 2 server farms
p = 0.1; % 10% of requests are HD
subscribers = 8000;
lambda = 1 / (24 * 7); % 1 request / week. lambda is requests/hour
lambda = lambda * subscribers;
invmiu = 90;
Ms = 2;
Mh = 5;
R = 10000;
N = 1000; % one month warm up
b_s = zeros(1,100);
b_h = zeros(1,100);
runs = 5;
W_limit = 100;
b_s_confidence = zeros(W_limit,2);
b_h_confidence = zeros(W_limit,2);
for W=1:W_limit
for lap=1:runs
[b_s(W,lap), b_h(W,lap)] = simulator2(lambda, p, invmiu, S, W, Ms, Mh, R, N);
end
[b_s_confidence(W,1), b_s_confidence(W,2)] = confidence_level(0.1, b_s(W,:), runs);
[b_h_confidence(W,1), b_h_confidence(W,2)] = confidence_level(0.1, b_h(W,:), runs);
fprintf('W:%.0f : %.6f +- %.6f || %.6f +- %.6f\n', W, b_s_confidence(W,1)*100, b_s_confidence(W,2)*100, b_h_confidence(W,1)*100, b_h_confidence(W,2)*100);
end
plot(1:W_limit,b_s_confidence(:,1), 1:W_limit, b_h_confidence(:,1));
dist = zeros(1,W_limit);
for i=1:W_limit
dist(i)= abs(b_s_confidence(i,1) - b_h_confidence(i,1));
end
find(dist==min(dist))
end
function ex3_c()
p = 0.2; % 20% HD requests
lambda = 1 / (24*7);
lambda = lambda * 17000; % 17000 subscribers
invmiu = 90;
Ms = 2;
Mh = 5;
R = 10000;
N = 1000;
S_limit = 6;
W_limit = 250;
runs = 40;
b_s_confidence = zeros(W_limit, S_limit);
b_s_confidence_error = zeros(W_limit, S_limit);
b_h_confidence = zeros(W_limit, S_limit);
b_h_confidence_error = zeros(W_limit, S_limit);
for S=1:S_limit
for W=0:W_limit
b_s = zeros(1,runs);
b_h = zeros(1,runs);
for lap=1:runs
[b_s(lap), b_h(lap)] = simulator2(lambda, p, invmiu, S, W, Ms, Mh, R, N);
end
[b_s_confidence(W+1,S), b_s_confidence_error(W+1,S)] = confidence_level(0.1, b_s, runs);
[b_h_confidence(W+1,S), b_h_confidence_error(W+1,S)] = confidence_level(0.1, b_h, runs);
fprintf('W %.0f S %.0f: %.5f || %.5f\n', W, S, b_s_confidence(W+1,S), b_h_confidence(W+1,S))
end
end
% compute worse case of the two streams
worse_case = zeros(size(b_s_confidence,1), size(b_s_confidence,2));
for i=1:size(b_s_confidence,1)
for j=1:size(b_s_confidence,2)
if b_s_confidence(i,j) >= b_h_confidence(i,j)
worse_case(i,j) = b_s_confidence(i,j);
else
worse_case(i,j) = b_h_confidence(i,j);
end
end
end
surf(1:S_limit, 0:W_limit, worse_case)
xlabel('Nr Servers');
ylabel('W reservation');
zlabel('Worst case');
%axis([1 S_limit 0 W_limit 0 0.5])
view(70,27)
grid on
% a = [2,2,3;0,2,5;1 2 3]
% [row,column]=find(a==min(min(a(a>0))))
[W_optimal, S_optimal] = find(worse_case==min(min(worse_case(worse_case>=0.001))))
menos_um_server=worse_case(:,1:5);
[W_optimal, S_optimal] = find(menos_um_server==min(min(menos_um_server)))
w=1
end