First few steps of simplex table for minimization

110 Views Asked by At

I saw this simple problem and realized I dont know how the simplex method solves this one in particular -

MIN c1x2 + c2x2 + c3x3
ST  x1 + x2 + x3 = 1

xi >= 0, ci >= 0 

(Where c3>c1>c2)

Typically with the simplex method we are supposed to make the objective function negative. Since c2 is the smallest, the answer is surely x2=1, x1=0, x3=0 but I dont know how to get there.

1

There are 1 best solutions below

4
On BEST ANSWER

In order to solve this model, we would need to add an artificial variable into the first constraint. For simplicity this is done via the Big-M approach as such:

$$\min z - c_1x_1-c_2x_2-c_3x_3-Ma_1=0$$ Subject to: $$x_1+x_2+x_3+a_1=1$$ $$x_1,x_2,x_3\ge0$$ $$c_2<c_1<c_3$$ Where $M$ is an arbitrarily "largest" number in $\Bbb R$. This would look like such in the tableau:

\begin{array} {|c|c|}\hline BV & z & x_1 & x_2 & x_3 & a_1 & RHS & RT \\ \hline z & 1 & -c_1 & -c_2 & -c_3 & -M & 0 & - \\ \hline ? & 0 & 1 & 1 & 1 & 1 & 1 & - \\ \hline \end{array}

Then we would have to solve for our basic variable $a_1$:

\begin{array} {|c|c|}\hline BV & z & x_1 & x_2 & x_3 & a_1 & RHS & RT \\ \hline z & 1 & M-c_1 & M-c_2 & M-c_3 & 0 & M & - \\ \hline a_1 & 0 & 1 & 1 & 1 & 1 & 1 & 1 \\ \hline \end{array}

Since $c_2$ is the smallest number to take away from $M$, the $x_2$ column is the largest $C^\pi_j$ greater than or equal to zero in the objective function row. Thus, the $x_2$ column is where we would pivot to get:

\begin{array} {|c|c|}\hline BV & z & x_1 & x_2 & x_3 & a_1 & RHS & RT \\ \hline z & 1 & -c_1+c_2 & 0 & -c_3+c_2 & -M+c_2& c_2 & - \\ \hline x_2 & 0 & 1 & 1 & 1 & 1 & 1 & - \\ \hline \end{array}

Since $c_2$ is smaller than $c_1$ and $c_3$, there are no more pivots available for Simplex to take. Thus, we are done.