我第一次尝试使用遗传算法。我犯了一个错误，遇到了麻烦。
使用无法识别的函数或变量‘A_UNFINTED’时出错(第393行)。微处理器_ga_Sanders(第10行)[BEST_X，BEST_OBJECT]=ga(@METPROCER_COST，NVARS，[]，lb，ub，@Constraint_Function_微处理器_Sanders，INTCON，OPTIONS)出错，原因：初始用户提供的非线性约束函数求值失败。
我需要一些帮助来评估这个问题。我已经为此工作了几个小时，但还是想不出来。

clc
clear
close all

options = optimoptions('ga','Display','iter','ConstraintTolerance',0);
NVARS=3; 
lb=[1, 0.0005, 0.0005];
ub=[30, 0.03, 0.03]; 
INTCON = [1];
[best_X,best_Objective] = ga(@microprocessor_cost,NVARS,[],[],[],[],lb,ub,@Constraint_function_microprocessor_Sanders,INTCON,options)

我将附加我在代码中引用的所有函数。
目标函数：

function [cost] = microprocessor_cost(X)
%function to determine the cost of an aluminum, rectangular heat sink
%   N is the number of fins
%   L is the length of the fins
%   t is the thickness of fins
%   b is the height of the base of the heat sink
w=0.0508; %w and r are hardcoded for the specific problem
r=0.0508;
b=0.002;
N=X(:,1);
L=X(:,2);
t=X(:,3);
volume_base=b*w*r;
volume_fins=L*w*t*N;
volume_total=volume_base+volume_fins;
cost=volume_total*849064; 
end

约束函数：

function [C, Ceq] = Constraint_function_microprocessor_Sanders(X)

%C is vector of inequality constraints. GA keeps it less than zero, i.e. C<0     and only solutions with C<0 are feasible
%Ceq is vecor of equality constraints (we have none). GA keeps it equal to zero, i.e. Ceq=0, and only solutions with Ceq=0 ARE FEASIBLE

Resistance = resistance_microprocessor(X);

%the following will make unrealistic designs infeasible
if A_unfinned<0
    Resistance = 1000;
end
Target_Resistance=50/15;
C = Resistance - Target_Resistance; % C<0, implies Resistance<Target_Resistance
Ceq = [];

end

阻力函数：

function [Resistance] = resistance_microprocessor(X)
%UNTITLED4 Summary of this function goes here
%   Detailed explanation goes here

%2in by 2in microprocessor - assuming that the width and length of the heat
%sink is equal to the width and length of the microprocessor
%2in = 0.0508m
w=0.0508; %m
l=0.0508; %m 2x2 inch microprocessor
fins=X(:,1);
L=X(:,2);
t=X(:,3);

%assuming no thermal grease between microprocessor and heat sink
contact_resistance_per_area=0.001; %K/(W*m^2)

Tamb=30+273; %K

v=0.000021; %m^2/s

Pr=0.7;

k_air=0.03; %W/(m*K)

%% Design deciscions
Tambient=293; %K, ambient air temperature
b=0.002; %m, height of base of heat sink
k=240; %W/(m*K), aluminum fins

%% Calculations

%calculated, (80+30)/2 = 55, (55+80)/2 = 42.5
Tfilm=42.5+273; %K

%chosen length of fins: 75 mm, which is the maximum length, Lc for vertical
%plate = L
Lc=0.075; %m, 

%solve for natural convection coefficient
h=function_get_convection_coefficient(Tfilm,Tamb,Lc,v,Pr,k_air); %W/m^2

%constraint - must be able to dissipate 15 W
q=15; %W

%thermal network calculations - thermal resistances

%contact resistance
area_transistor=w*l;
R_contact=contact_resistance_per_area*area_transistor;

%resistance - conduction through base
A_base=l*w;
R_base=b/(k*A_base);

%fin thermal resistance - convection
n=function_get_fin_efficiency(h,L,t,w,k); %fin efficiency using function
surf_area_fins=fins*2*(L*w); %ignoring the convection through the end and through top and bottom
R_fins=1/(h*surf_area_fins*n);

%unfinned thermal resistance
A_unfinned=(l-(fins*t))*w;
R_unfinned=1/(h*A_unfinned);

%overall thermal resistance of system
Resistance=R_contact+R_base+(1/(1/R_fins)+(1/R_unfinned));

end