# Expressing a chain of boolean if-then with logical ANDs using MIP

How to express a chain of boolean If-then as MIP such as:

If $$(x_{10} \ge b_1$$ and $$x_{11} \le b_1)$$ AND $$(x_{20} \ge b_2$$ and $$x_{21} \le b_2)$$... AND... then $$y_1 = 1$$ else $$y_1 = 0$$.

So basically, if constants are between values of different $$x_0$$ and $$x_1$$ (lower and upper bounds for every constant) then $$y = 1$$, if at least one constant is out of bound then $$y = 0$$. I need to maximize sum of all $$y$$.

• AFAIK, some commercial solvers like CPLEX or GUROBI have had such capability to write such constraints directly in the static model. Would you see this link? Jan 21 '20 at 5:40

Because you are maximizing the sum of $$y$$, as long as each $$y$$ appears in no other constraints, you can get by with enforcing only one direction of the implication, namely $$y_1=1\implies (x_{10}\ge b_1 \land x_{11}\le b_1 \land \dots).$$ You can do this with one big-M constraint for each inequality, where $$\ell$$ and $$u$$ are (constant) lower and upper bounds on $$x$$: \begin{align} b_1-x_{10}&\le (b_1-\ell_{10})(1-y_1)\\ x_{11}-b_1&\le (u_{11}-b_1)(1-y_1)\\ b_2-x_{20}&\le (b_2-\ell_{20})(1-y_1)\\ x_{21}-b_2&\le (u_{21}-b_2)(1-y_1)\\ &\dots \end{align}