Lousy Big-M relaxation with implies

[somil miglani]

Dear all, 

I am trying to use the Implies function in yalmip and I get the following warning:

"You have unbounded variables in IMPLIES leading to a lousy big-M relaxation"

Through the older posts on the same issue I learnt that the variables involved should have explicit bounds. I have tried to put the bounds but the warning does not go away. 

Besides, the solution progresses very slow. Could it be because of the implies conditions used? 

Following is the scheme I am trying to implement;

A: Continuous variable (5 X 100)

B: Continuous variable (1 X 100)

y: binary variable (5 X 100)

Constraints:

C1: sum(y, 1) <= 1

C2: A >= y*Amin

C3: A <= y*Amax

C4: B <= Amin

C5: B >= Amax

C6: implies(sum(y, 1),  sum(A, 1) == B)

I have tried to change the bound constraints to sum(A, 1) >= sum(y, 1)*Amin and similar for upper bound.  The warning still stays. 

Any help is appreciated. 

Thanks

Somil 

[Johan Löfberg]

A is not explicitly bounded