Tdata 1x26 208 double
Vdata 1x26 208 double
Tdata =
-40 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50 55
60 65 70 75 80 85
Vdata=
0.1251 0.1203 0.1161 0.1124 0.1094 0.1068 0.1049 0.1034 0.1026 0.1023 0.1026 0.1034
0.1049 0.1068 0.1093 0.1124 0.1161 0.1203 0.1251 0.1304 0.1364 0.1428 0.1499 0.1574
0.1656 0.1743
%% Temperature and Voltage Data
Tdata = -40:5:85;
Vdata = 1.026E-1 + -1.125E-4 * Tdata + 1.125E-5 * Tdata.^2;
%% Plot the Desired Curve
plot(Tdata,Vdata,'-*');
title('Target Curve','FontSize',12);
xlabel('Temperature (^oC)'); ylabel('Voltage (V)')
load('StandardComponentValues.mat')
Res 70x1 560 double
ThBeta 9x1 72 double
ThVal 9x1 72 double
ThBeta =
2750
3680
3560
3620
3528
3930
3960
4090
3740
ThVal =
50
220
1000
2200
3300
4700
10000
22000
33000
Res =
300
330
360
390
430
470
510
560
620
680
750
820
910
1000
1100
1200
1300
1500
1600
1800
2000
2200
2400
2700
3000
3300
3600
3900
4300
4700
5100
5600
6200
6800
7500
8200
9100
10000
11000
12000
13000
15000
16000
18000
20000
22000
24000
27000
30000
33000
36000
39000
43000
47000
51000
56000
62000
68000
75000
82000
91000
100000
110000
120000
130000
150000
160000
180000
200000
220000
Vnew = voltageCurve(Tdata, [2 2 2 2 1 1], Res, ThVal, ThBeta);
%% Add New Curve to Plot
hold on; plot(Tdata,Vnew,'-or');
Vnew =1x26 double
0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500 0.5500
%%
% We would like to find the optimal indices that result in a temperature
% curve closest to our desired curve. Our initial guess of [2 2 2 2 1 1]
% did not yield a good fit, and there are 70*70*70*70*9*9 = 1944810000
% different possible combinations for this circuit. We will use an
% optimization routine, the Genetic Algorithm, to find the optimal indices
% for our problem. The Genetic Algorithm allows us to constrain values to
% be integers, which is necessary for this problem since indices into a
% vector must be integers.
%% Bounds on our Vector of Indices
lb = [1 1 1 1 1 1];
ub = [70 70 70 70 9 9];
%% Constrain All 6 Variables to be Integers
intCon = 1:6;
custOutput = @(a1,a2,a3)ThOptimPlot(a1,a2,a3,Tdata,Vdata,Res,ThVal,ThBeta);
options = gaoptimset('CrossoverFrac',0.5,'PopulationSize',100,...
'StallGen',125,'Generations',150,...
'PlotFcns',@gaplotbestf,'OutputFcns',custOutput);
%% Run the Genetic Algorithm
% Note: The Genetic Algorithm is based on stochastic methods, meaning that
% we are not guaranteed to find the same solution every time. To reproduce
% the exact results from the video, run "rng(45)" at the Command Line to
% set the same seed for the random number generator.
[xOpt,fVal] = ga(@(x)objectiveFunction(x,Res,ThVal,ThBeta,Tdata,Vdata),...
6,[],[],[],[],lb,ub,[],intCon,options);
%% Inspect the Solution Vector to see that All Values are Integers
disp('Integer Solution Returned by GA:')
disp(xOpt)
