zeros populating most of my output vector

I am trying to iterate over a changing variable in a vector so that I use said variable in a function. In this program I want to keep my output between 0.01 and 0.008. To achieve this I have to vary another different variable within the script using if statments. My functions are all working correctly so i am just wondering where i am going wrong in my process?

clear all
close all
clc
no_of_calls = [6 3 2 2 5 29 343 548 535 500 535 593 600 583 571 553 458 316 206 123 64 28 13]; %init no of calls
no_of_runs = 1;
no_of_channels = 46;
for i = 1:23
    GOS = zeros(length(no_of_calls(13)), no_of_runs); % create GOS vector
    
    no_of_channels
    
    
    
    GOS(i) = part_3(no_of_calls(i), no_of_runs, no_of_channels);
    
    while GOS(i) > 0.01
        
        GOS(i) = part_3(no_of_calls(i), no_of_runs, no_of_channels);         
        no_of_channels
        
        if GOS(i) > 0.01
            no_of_channels = no_of_channels+1;
            if i ~= 1
                i = i - 1;
            else
                i = 1;
            end
        end
        
    end
    
    while GOS(i) < 0.008
        
        GOS(i) = part_3(no_of_calls(i), no_of_runs, no_of_channels); 
        no_of_channels
        
        if GOS(i) < 0.008
            no_of_channels = no_of_channels-1;
            if i ~= 1
                i = i - 1;
            else
                i = 1;
            end
        end
        
    end
    
    
end

[David Hill]

You should be able to greatly simplify. Show us your part_3 function. I see lots of problems with the above code.

[Cormac Carr]

function GOS = part_3(no_of_calls,no_of_runs,no_of_channels )
% set up variables
CallTime = zeros(1,no_of_runs);
n = ones(1,no_of_calls).*no_of_channels;
E_1 = zeros(no_of_runs, no_of_calls);
call_timeDIST= makedist('Uniform');
call_lengthDIST = makedist('lognormal','mu',0.69,'sigma',0.44);
%A_0 = zeros(no_of_runs, no_of_calls);
%call_lengthDIST = makedist('poisson','lambda',3);
CallLength = zeros(no_of_runs,no_of_calls);
for i = 1:no_of_calls
    for j = 1:no_of_runs
        % Create distribution for calls in time
        CallTime(j) = random(call_timeDIST);
        % Create distribution for call length
        CallLength(j,i) = random(call_lengthDIST);
        % calculate A0
        A_0       = no_of_calls.*(CallLength./60);
        % calculate E_1 using Erlings-B Formula
        E_1(j,i) = part_2(A_0(j,i), n(i));
    end
end
% Find Mean for number of iterations
GOS = mean(E_1,2,"omitnan");
r=rand(no_of_runs,no_of_calls);%generate randomn uniform array
dropped_calls = nnz(r<E_1);%compare the two arrays, and count the number of instances.
CT = sort(CallTime)';
xaxis = 1:no_of_runs;

[David Hill]

You are not providing all the pieces to be able to run. Where is the function part_2?

[Cormac Carr]

function E_1 = part_2(A_0,n) 
dnom = zeros(1,n);
num1 = zeros(1,n);
E_1 = zeros(1,n);
for i = 1:n
    S(i) = (A_0^i)/factorial(i); %% Create vector of 
    %                               values to be summed 
    %                               to create denominator
    dnom = cumsum(S);            %% Create denominator
    
end
for n = 1:n
    num1(n) = ((A_0^n)/factorial(n)); %% loop through values
    %  values of n for numerator
    %  createing vector of numerators
    %  for specified A0
end
%  disp("NValue")
%  n
for n = 1:n
    %     disp("here")
    E_1 = num1(n)./dnom(n);                        % Complete formula for answer
    
end

[Jan]

@Cormac Carr: Please use the tools to format code as code. Thanks.

What is the purpose of length(no_of_calls(13)) ? This is 1 in all cases. no_of_calls(13) is a scalar, so its length is 1.

A nicer replacement of:

if i ~= 1
    i = i - 1;
else
    i = 1;
end  

is:

i = max(i - 1, 1);

Answer 1

[David Hill]

You will start running into problems with higher no_of_calls because the no_of_channels will increase and factorial numbers will become too large.

no_of_calls = [6 3 2 2 5 29];
no_of_runs = 1;
GOS=zeros(1,length(no_of_calls));
no_of_channels=ones(size(GOS))*46;
for i = 1:length(no_of_calls)
    GOS(i) = part_3(no_of_calls(i), no_of_runs, no_of_channels(i));
    while GOS(i)>0.01 || GOS(i)<.008
        if GOS(i)>.01
           no_of_channels(i) = no_of_channels(i)+1;
        else
           no_of_channels(i) = no_of_channels(i)-1;
        end
        GOS(i) = part_3(no_of_calls(i), no_of_runs, no_of_channels(i));
    end
end

GOS

GOS = 1×6

0.0100 0.0099 0.0100 0.0100 0.0098 0.0097

no_of_channels

no_of_channels = 1×6

108 104 102 104 108 145

function E_1 = part_2(A_0,n) 
r=(1:n)';
num1 = A_0.^r./factorial(r);
dnom = cumsum(num1,1);  
E_1 = (num1./dnom)';                       
end

function GOS = part_3(no_of_calls,no_of_runs,no_of_channels)
n = no_of_channels;
call_lengthDIST = makedist('lognormal','mu',0.69,'sigma',0.44);
CallLength = random(call_lengthDIST,no_of_runs,no_of_calls);
A_0 = no_of_calls*CallLength/60;
E_1 = part_2(A_0, n);
GOS = mean(mean(E_1,2,"omitnan"));
end