Using implies command

[MY]

Hi,

I would like to do the following logic calculations with implies command but not sure how to do it.

My original if statement is as follows:

if t>1  % t is time

    tm = num2str( t-1 );

    eval(['Rtm = R_kl' tm ';'])  % R_kl=binvar(n,1)

    eval(['Rt = R_kl' tt ';'])

if Rtm==Rt

Obj(t) = Obj(t)

else

 Obj(t) = Obj(t)+1000;

end

This is how I do it by implies:

if t>1

    tm = num2str( t-1 );

    eval(['Rtm = R_kl' tm ';'])

    eval(['Rt = R_kl' tt ';'])

implies(Rtm==Rt,Obj(t)==Obj(t))

else 

    Obj(t) = Obj(t)+1000;

end

[Johan Löfberg]

the constraint obj(t) == obj(t) makes no sense (as it always is true), and obj(t) = obj(t)+100 is not well defined (what is obj(t) from the beginning?)

and the code you start with is not correct as there is a missing end, so it is very unclear how the logic is supposed to be

[MY]

Let me try to make it more clear.

for t=1:T  % t is time
 tt = num2str( t );
eval([' X' tt ' = sdpvar(n,n)'])
eval([ 'R_kl' tt ' = binvar(9,1)']);
eval([' Obj(t)= trace(A(t)*X' tt ')']) % A is defined matrix

if t>1  
    tm = num2str( t-1 );
    eval(['Rtm = R_kl' tm ';'])
    eval(['Rt = R_kl' tt ';'])

if Rtm~=Rt
 Obj(t) = Obj(t)+1000;

end
else
end
end

How I can do this y using implies?

Thanks,

[Johan Löfberg]

looks like you want to do

for t=1:T  % t is

time
    X{t} = sdpvar(n,n);
    R{t} = binvar(9,1);
       
    if t>1
        sdpvar z
        cases = binvar(2,1);
        Model = [Model, implies(cases(1), [R{t}==R{t-1}, z == trace(A(t)*X{t})])
                        implies(cases(2), [sum(abs(R{t}-R{t-1}))>= 1/2, z == trace(A(t)*X{t})+100])                        
                        sum(cases) == 1];
      Obj = [Obj z];
    else
      Obj = trace(A(t)*X{t});
    end
end

although expression A(t)*X{t} looks dubious. One would express this is supposed to be the inner product of two matrices, but A(t) is a scalar...

it also looks strangely unconstrained in X, i.e. constraints missing

[MY]

I tried your suggestion, but not able to get feasible solution.

I'm wondering if it is possible for me to send my code to your personal email, so you may have a look at it?

Thanks,

[Johan Löfberg]

sure

[Evgeny Schnittmann]

Hello Mr. Löfberg,

at the Moment im trying to restrict certain cases in my yalmip-optimization. Im trying to do it with the implies function. My question is, if it is possible to formulate a Logical "&" in that context.

How can I for example translate the following Code into an implies-constraint?

if a && b

    c=1;

end

My solution for now is:

if a

   Constraints = [Constraints, implies(b,c)]

end

Is there a better alternative?

Thank you in advance!

[Johan Löfberg]

So a is constant? Then that is the most natural model yes

[Evgeny Schnittmann]

Yes a is from 1 to horizon (unit commitment) and b is of the type sdpvar. Anyway, thank you very much!

[Johan Löfberg]

and c is binary too? Otherwise you must write implies(b,c==1)

and of course, if both b and c are binary, you are simply saying c>=b

[Johan Löfberg]

...assuming b binary too of course

[Evgeny Schnittmann]

You are right, i forgot the second "=" (not in my Code though). no, c and b are both continuous.

[Johan Löfberg]

if b and c are continuous, then b implies c does not make any numerical sense. So if b = 10^-40, you want to avoid c=0 but you are ok with c = 10^-32