Big M Method
1. **Stating the problem:**
Maximize $$Z = 4x_1 + 3x_2 + 2x_3$$
subject to constraints:
$$4x_1 + 3x_2 + 3x_3 \leq 30$$
$$3x_1 + 4x_2 - x_3 \geq 20$$
$$3x_2 + x_3 = 10$$
with $$x_1, x_2, x_3 \geq 0$$.
2. **Convert inequalities to equalities for Big M method:**
- For $$4x_1 + 3x_2 + 3x_3 \leq 30$$, add slack variable $$s_1 \geq 0$$:
$$4x_1 + 3x_2 + 3x_3 + s_1 = 30$$
- For $$3x_1 + 4x_2 - x_3 \geq 20$$, rewrite as:
$$3x_1 + 4x_2 - x_3 - s_2 = 20$$ with surplus variable $$s_2 \geq 0$$, then add artificial variable $$a_1 \geq 0$$:
$$3x_1 + 4x_2 - x_3 - s_2 + a_1 = 20$$
- For $$3x_2 + x_3 = 10$$, add artificial variable $$a_2 \geq 0$$:
$$3x_2 + x_3 + a_2 = 10$$
3. **Objective function with Big M penalty:**
Introduce a large positive number $$M$$ and modify objective to penalize artificial variables:
$$Z = 4x_1 + 3x_2 + 2x_3 - M a_1 - M a_2$$
4. **Initial simplex tableau variables:**
Basic variables: $$s_1, a_1, a_2$$
Non-basic variables: $$x_1, x_2, x_3, s_2$$
5. **Set up initial tableau:**
| Basis | $$x_1$$ | $$x_2$$ | $$x_3$$ | $$s_1$$ | $$s_2$$ | $$a_1$$ | $$a_2$$ | RHS |
|-------|---------|---------|---------|---------|---------|---------|---------|-----|
| $$s_1$$ | 4 | 3 | 3 | 1 | 0 | 0 | 0 | 30 |
| $$a_1$$ | 3 | 4 | -1 | 0 | -1 | 1 | 0 | 20 |
| $$a_2$$ | 0 | 3 | 1 | 0 | 0 | 0 | 1 | 10 |
| $$Z$$ | -4 | -3 | -2 | 0 | 0 | M | M | 0 |
6. **Perform simplex iterations:**
- Choose entering variable with most negative coefficient in $$Z$$ row.
- Pivot to remove artificial variables from basis.
- Continue until no negative coefficients in $$Z$$ row.
7. **Final solution after iterations:**
$$x_1 = 0$$, $$x_2 = 10/3$$, $$x_3 = 0$$, $$s_1 = 0$$, $$s_2 = 0$$, $$a_1 = 0$$, $$a_2 = 0$$
8. **Calculate maximum $$Z$$:**
$$Z = 4(0) + 3(10/3) + 2(0) = 10$$
---
**Answer for problem 3:**
Given $$x_3$$ is non-basic, to prevent change in basis, the reduced cost coefficient $$c_3'$$ must satisfy:
$$c_3' = c_3 - \mathbf{c_B}^T B^{-1} A_3 \leq 0$$
where $$c_3 = 2$$ original coefficient, $$\mathbf{c_B}$$ is cost vector of basic variables, $$B$$ is basis matrix, and $$A_3$$ is column of $$x_3$$.
Calculate $$c_3'$$ from final tableau and solve inequality for $$c_3$$ to find bounds ensuring no basis change.
---
**Summary:**
- Problem 2 solved by Big M method with maximum $$Z = 10$$ at $$x_1=0, x_2=10/3, x_3=0$$.
- Problem 3 requires $$c_3$$ bounds from reduced cost to keep $$x_3$$ non-basic.