addRule throws "Do not use a rule keyword as a variable name."

I get the error on line 44 of my code: fis = addRule(fis,all); I've tried changing the name of the variable name (all) to other names, and the fis. I've removed spaces from my variables or renamed them completely.

% Create a Fuzzy Inference System (FIS) with the name "IFR"
fis = mamfis('Name', 'IFR');
% Add a variable "input" to the FIS
fis = addInput(fis, [0 200], 'Name', 'MAP_mm');
% Add membership functions for variable "input"
fis = addMF(fis, "MAP_mm", "trapmf", [0 0 55 75], 'Name',"Low");   
fis = addMF(fis, "MAP_mm", "trapmf", [55 75 100 120], 'Name',"Normal");
fis = addMF(fis, "MAP_mm", "trapmf", [100 120 200 200], 'Name',"High");    
% HOU trapezoidal
fis = addInput(fis, [0 200], 'Name', 'HOU_ml-hr');
% HOU Low trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [0 0 30 40], 'Name',"Low");
% HOU Normal trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [30 40 100 125], 'Name',"Normal");
% HOU High trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [100 125 200 200], 'Name',"High");
% Add output variable to the FIS
fis = addOutput(fis, [0 2000], 'Name', 'IFR_ml-hr');
% Add membership functions for output variable
fis = addMF(fis, "IFR_ml-hr", "trapmf", [0 0 60 100], 'Name',"Low");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [0 100 200 400], 'Name',"Maintain");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [200 400 600 800], 'Name',"Moderate");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [600 800 1000 1500], 'Name',"High");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [1000 1500 2000 2000], 'Name',"Very_High");
% Create Rules for the FIS
r1 = "If HOU_ml-hr == Low & MAP_mm == Low => IFR_ml-hr == Very_High";
r2 = "If HOU_ml-hr == Normal & MAP_mm == Low => IFR_ml-hr == High";
r3 = "If HOU_ml-hr == High & MAP_mm == Low => IFR_ml-hr == Moderate";
r4 = "If HOU_ml-hr == Low & MAP_mm == Normal => IFR_ml-hr == Moderate";
r5 = "If HOU_ml-hr == Low & MAP_mm == High => IFR_ml-hr == Low";
r6 = "If HOU_ml-hr == Normal & MAP_mm == Normal => IFR_ml-hr == Maintain";
r7 = "If HOU_ml-hr == Normal & MAP_mm == High => IFR_ml-hr == Low";
r8 = "If HOU_ml-hr == High & MAP_mm == Normal =>IFR_ml-hr == Maintain";
r9 = "If HOU_ml-hr == High & MAP_mm == High => IFR_ml-hr == Low";
all = [r1 r2 r3 r4 r5 r6 r7 r8 r9];
% Add rules and enable rule viewing/debugging
fis = addRule(fis,all);

Error using FuzzyInferenceSystem/addRule (line 1153)
Do not use a rule keyword as a variable name.

% Evaluate the FIS for inputs [MAP HOU]
sampleInput = [110 120; 60 25; 30 150; 180 90];
out = evalfis(fis, sampleInput);
% Plot membership functions
subplot(3,1,1)
plotmf(fis, 'input', 1)
subplot(3,1,2)
plotmf(fis, 'input', 2)
subplot(3,1,3)
plotmf(fis, 'output',3)

Answer 1

[Walter Roberson]

Rewriting the rules as words works.

% Create a Fuzzy Inference System (FIS) with the name "IFR"
fis = mamfis('Name', 'IFR');
% Add a variable "input" to the FIS
fis = addInput(fis, [0 200], 'Name', 'MAP_mm');
% Add membership functions for variable "input"
fis = addMF(fis, "MAP_mm", "trapmf", [0 0 55 75], 'Name',"Low");   
fis = addMF(fis, "MAP_mm", "trapmf", [55 75 100 120], 'Name',"Normal");
fis = addMF(fis, "MAP_mm", "trapmf", [100 120 200 200], 'Name',"High");    
% HOU trapezoidal
fis = addInput(fis, [0 200], 'Name', 'HOU_ml-hr');
% HOU Low trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [0 0 30 40], 'Name',"Low");
% HOU Normal trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [30 40 100 125], 'Name',"Normal");
% HOU High trapezoidal
fis = addMF(fis, "HOU_ml-hr", "trapmf", [100 125 200 200], 'Name',"High");
% Add output variable to the FIS
fis = addOutput(fis, [0 2000], 'Name', 'IFR_ml-hr');
% Add membership functions for output variable
fis = addMF(fis, "IFR_ml-hr", "trapmf", [0 0 60 100], 'Name',"Low");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [0 100 200 400], 'Name',"Maintain");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [200 400 600 800], 'Name',"Moderate");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [600 800 1000 1500], 'Name',"High");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [1000 1500 2000 2000], 'Name',"Very_High");
% Create Rules for the FIS
r1 = "If HOU_ml-hr is Low and MAP_mm is Low then IFR_ml-hr is Very_High";
r2 = "If HOU_ml-hr is Normal and MAP_mm is Low then IFR_ml-hr is High";
r3 = "If HOU_ml-hr is High and MAP_mm is Low then IFR_ml-hr is Moderate";
r4 = "If HOU_ml-hr is Low and MAP_mm is Normal then IFR_ml-hr is Moderate";
r5 = "If HOU_ml-hr is Low and MAP_mm is High then IFR_ml-hr is Low";
r6 = "If HOU_ml-hr is Normal and MAP_mm is Normal then IFR_ml-hr is Maintain";
r7 = "If HOU_ml-hr is Normal and MAP_mm is High then IFR_ml-hr is Low";
r8 = "If HOU_ml-hr is High and MAP_mm is Normal then IFR_ml-hr is Maintain";
r9 = "If HOU_ml-hr is High and MAP_mm is High then IFR_ml-hr is Low";
all = [r1 r2 r3 r4 r5 r6 r7 r8 r9];
% Add rules and enable rule viewing/debugging
fis = addRule(fis,all);
% Evaluate the FIS for inputs [MAP HOU]
sampleInput = [110 120; 60 25; 30 150; 180 90];
out = evalfis(fis, sampleInput);
% Plot membership functions
subplot(3,1,1)
plotmf(fis, 'input', 1)
subplot(3,1,2)
plotmf(fis, 'input', 2)
subplot(3,1,3)
plotmf(fis, 'output',1)

●

[Michael]

Thanks! I get another error when I run the code now, though i believe it is unrelaed. I get"Access violation - no RTTI data!"

Thank you for the help! I really appreciate it.

[Walter Roberson]

There used to be a bug report about that message, https://www.mathworks.com/support/bugreports/1465379?s_tid=answers_rc2-2_p5_MLT that was related to savefig()

Answer 2

[Sam Chak]

Hi @Michael

If you prefer to use the symbolic approach (==, =>) rather than the linguistic approach (If–Then) as demonstrated by @Walter Roberson, do not use the reserved fuzzy rule-related terms such as "If" and "Then". In your case, when you specified "If HOU_ml-hr == Low", MATLAB interprets this as if you had inadvertently used the phrase "If HOU_ml-hr" as a single variable name. I have made the corrections below, as well as slightly "beautified" the plots to prevent overlapping linguistic names.

If you like the explanation and the enhanced plots, please consider voting 👍 for this Answer.

% Create a Fuzzy Inference System (FIS) with the name "IFR"
fis = mamfis('Name', 'IFR');

% Input variable "MAP"
fis = addInput(fis, [0 200], 'Name', 'MAP_mm');
fis = addMF(fis, "MAP_mm",    "trapmf", [   0    0   55   75], 'Name', "Low");   
fis = addMF(fis, "MAP_mm",    "trapmf", [  55   75  100  120], 'Name', "Normal");
fis = addMF(fis, "MAP_mm",    "trapmf", [ 100  120  200  200], 'Name', "High");   

% Input variable "HOU"
fis = addInput(fis, [0 200], 'Name', 'HOU_ml-hr');
fis = addMF(fis, "HOU_ml-hr", "trapmf", [   0    0   30   40], 'Name', "Low");
fis = addMF(fis, "HOU_ml-hr", "trapmf", [  30   40  100  125], 'Name', "Normal");
fis = addMF(fis, "HOU_ml-hr", "trapmf", [ 100  125  200  200], 'Name', "High");

% Output variable "IFR"
fis = addOutput(fis, [0 2000], 'Name', 'IFR_ml-hr');
fis = addMF(fis, "IFR_ml-hr", "trapmf", [   0    0   60  100], 'Name', "Low");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [   0  100  200  400], 'Name', "Maintain");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [ 200  400  600  800], 'Name', "Moderate");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [ 600  800 1000 1500], 'Name', "High");
fis = addMF(fis, "IFR_ml-hr", "trapmf", [1000 1500 2000 2000], 'Name', "Very_High");

% Create Rules for the FIS
% r1 = "If HOU_ml-hr==Low    & MAP_mm==Low    => IFR_ml-hr==Very_High";
% r2 = "If HOU_ml-hr==Normal & MAP_mm==Low    => IFR_ml-hr==High";
% r3 = "If HOU_ml-hr==High   & MAP_mm==Low    => IFR_ml-hr==Moderate";
% r4 = "If HOU_ml-hr==Low    & MAP_mm==Normal => IFR_ml-hr==Moderate";
% r5 = "If HOU_ml-hr==Low    & MAP_mm==High   => IFR_ml-hr==Low";
% r6 = "If HOU_ml-hr==Normal & MAP_mm==Normal => IFR_ml-hr==Maintain";
% r7 = "If HOU_ml-hr==Normal & MAP_mm==High   => IFR_ml-hr==Low";
% r8 = "If HOU_ml-hr==High   & MAP_mm==Normal => IFR_ml-hr==Maintain";
% r9 = "If HOU_ml-hr==High   & MAP_mm==High   => IFR_ml-hr==Low";

r1 = "HOU_ml-hr==Low    & MAP_mm==Low    => IFR_ml-hr==Very_High";
r2 = "HOU_ml-hr==Normal & MAP_mm==Low    => IFR_ml-hr==High";
r3 = "HOU_ml-hr==High   & MAP_mm==Low    => IFR_ml-hr==Moderate";
r4 = "HOU_ml-hr==Low    & MAP_mm==Normal => IFR_ml-hr==Moderate";
r5 = "HOU_ml-hr==Low    & MAP_mm==High   => IFR_ml-hr==Low";
r6 = "HOU_ml-hr==Normal & MAP_mm==Normal => IFR_ml-hr==Maintain";
r7 = "HOU_ml-hr==Normal & MAP_mm==High   => IFR_ml-hr==Low";
r8 = "HOU_ml-hr==High   & MAP_mm==Normal => IFR_ml-hr==Maintain";
r9 = "HOU_ml-hr==High   & MAP_mm==High   => IFR_ml-hr==Low";

all = [r1 r2 r3 r4 r5 r6 r7 r8 r9];
fis = addRule(fis, all);

% Evaluate the FIS for inputs [MAP HOU]
% sampleInput = [110 120; 60 25; 30 150; 180 90];
% out = evalfis(fis, sampleInput);

% Plot membership functions
figure
subplot(3,1,1)
plotmf(fis, 'input', 1)
ax1 = gca;
delete(findall(ax1, 'type', 'text'))
text( 25, 1.2, 'Low');
text( 77, 1.2, 'Normal');
text(155, 1.2, 'High');
ylim([-0.2, 1.4])
ylabel({'$\mu$'},    'interpreter', 'latex')
xlabel({'MAP / mm'}, 'interpreter', 'latex')

subplot(3,1,2)
plotmf(fis, 'input', 2)
ax2 = gca;
delete(findall(ax2, 'type', 'text'))
text( 10, 1.2, 'Low');
text( 60, 1.2, 'Normal');
text(155, 1.2, 'High');
ylim([-0.2, 1.4])
ylabel({'$\mu$'},       'interpreter', 'latex')
xlabel({'HOU / ml-hr'}, 'interpreter', 'latex')

subplot(3,1,3)
% plotmf(fis, 'output', 3)    % there is no Output number 3
plotmf(fis, 'output', 1)
ax3 = gca;
delete(findall(ax3, 'type', 'text'))
text(   0, 1.2, 'Low');
text(  40, 0.8, 'Maintain');
text( 365, 1.2, 'Moderate');
text( 830, 0.8, 'High');
text(1600, 1.2, 'Very High');
ylim([-0.2, 1.4])
ylabel({'$\mu$'},       'interpreter', 'latex')
xlabel({'IFR / ml-hr'}, 'interpreter', 'latex')

sgtitle("Membership functions for MAP, HOU, and IFR", 'fontsize', 12)

●

figure
opt = gensurfOptions('NumGridPoints', 51);
gensurf(fis, opt);
ax4 = gca;
delete(findall(ax4, 'type', 'text'))
xlabel({'MAP / mm'},    'interpreter', 'latex')
ylabel({'HOU / ml-hr'}, 'interpreter', 'latex')
zlabel({'IFR / ml-hr'}, 'interpreter', 'latex')
[az, el] = view()

az = -37.5000

el = 30

view(az+180, el)
title("Surface of Human Decision for IFR")

[Michael]

Love this answer! I ended up just using a generic scatter plot. As for my previous error, "Access violation - no RTTI data!", it was just Mathworks problem. I had to install updates and restart then the error went away. Thank you for the help!

[Walter Roberson]

Good catch!

Answer 3

[Michael]

"Access violation - no RTTI data!", it was just Mathworks problem. I had to install updates and restart then the error went away.