What instruction code can i use if i want to update the former parameter value by using the latter parameter value ?

I write a code ,and the latter part of code will produce a new parameter value to update the parameter value in the former part of code,what Instruction code can i use to realize what i want ?

For example, part 1 code will produce A,B these two values,

part 2 code will use the A,B value to find the C,D values

part 3 code will let A=C,B=D,and "return to part 2 to find the new C and D value,here we call the new C,D is E,F "

So we find the example above has to update the A,B values,so what instruction code can i use? Is there any example?

Answer 1

[Walter Roberson]

%part 1
A = randi([-9 9]);
B = randi([-9 9]);

for pass_number = 1 : 2
    if pass_number == 1
        %part 2
        C = A.^2 - B;
        D = A.^B;
    else
        E = A.^2 - B;
        F = A.^B;
        break;
    end
    %part 3
    A = C;
    B = D;
end

The break is there because there is no instruction to continue on to part 3 after having found the E and F the second time through part 2.

[yang-En Hsiao]

so if i delete the

break

,Will the E and F be the newest A and B and keep calculating and updating A and B?

what is the true meaning about

for pass_number = 1 : 2
    if pass_number == 1

i mean,why do i have to write this line code ? does it mean that the loop can only do two times?what the loop can do lots of time ,but it will stop when satisfy some of conditions?

[Walter Roberson]

it means the loop would only process two times. you only defined what should happen for two times.

E and F will not be the newest A and B. You defined A and B as being updated from C and D but you only set C and D once, the first time you go through part 2.

[Walter Roberson]

What is the reason for wanting C and D to be different variables than E and F ?

A common technique in code is

initialize one or more variables
compute a value based on the variables
compute an error measure
while the error measure is larger than you want
    compute new variables based on the current values of the variables
    compute a value based upon the new variables
    compute an error measure
    update the current variables based upon the new variables
end

Perhaps the most common use of this technique with two input variables is in doing a bisection search, possibly for the purpose of finding zeros of a function.