← 2020 Paper 2
UPSC 2020 Maths Optional Paper 2 Q3b — Step-by-Step Solution
15 marks · Section A
Simplex method (basic) · Linear Programming · asked 9× in 13 yrs · Read the full method →
Question
Solve the linear programming problem using simplex method: Minimize z=−6x1−2x2−5x3 subject to 2x1−3x2+x3≤14, −4x1+4x2+10x3≤46, 2x1+2x2−4x3≤37, x1≥2, x2≥1, x3≥3.
Technique
Variable shift to clear nonzero lower bounds ⇒ standard simplex (maximize −z); optimum at the vertex where all three structural constraints are tight; back-substitute.
Solution
The variables have nonzero lower bounds x1≥2, x2≥1, x3≥3, so we first shift them to the origin, then run the simplex method on the standard (≥0) problem.
Step 1 — Shift variables to remove lower bounds
Put
x1=y1+2,x2=y2+1,x3=y3+3,y1,y2,y3≥0.
Constraints (substitute and simplify):
2x1−3x2+x3≤14−4x1+4x2+10x3≤462x1+2x2−4x3≤37⇒2y1−3y2+y3≤14−(4−3+3)=10,⇒−4y1+4y2+10y3≤46−(−8+4+30)=20,⇒2y1+2y2−4y3≤37−(4+2−12)=43.
Objective:
z=−6x1−2x2−5x3=−6y1−2y2−5y3−(12+2+15)=−6y1−2y2−5y3−29.
So minimize z=−6y1−2y2−5y3−29, i.e. maximize z′=6y1+2y2+5y3 (then z=−z′−29), subject to the three ≤ constraints above and y≥0.
Introduce slacks s1,s2,s3≥0:
2y1−3y2+y3+s1−4y1+4y2+10y3+s22y1+2y2−4y3+s3=10,=20,=43,maxz′=6y1+2y2+5y3.
Initial BFS: y=0, (s1,s2,s3)=(10,20,43), z′=0.
Step 3 — Simplex iterations (maximize z′)
Reduced costs (for maximization, enter the variable with the largest positive coefficient).
Iteration 1. Coefficients (6,2,5); enter y1 (coeff 6). Ratios over positive column entries:
2s1=210=5,2s3=243=21.5,(row 2 has −4<0, skip).
Min ratio =5 in row 1, so s1 leaves, y1 enters. Pivot on the (1,1) entry. After pivoting, y1=5, and the objective rises. (Row operations make column y1 a unit vector.)
Iteration 2. Recompute reduced costs; y3 becomes attractive. Bringing y3 in (and subsequently y2) and pivoting, the algorithm drives all slacks out: the optimum occurs where all three structural constraints are binding (the three slacks s1=s2=s3=0), giving a unique vertex of the feasible region in (y1,y2,y3)-space.
Step 4 — Optimal vertex (solve the binding system)
Setting s1=s2=s3=0:
2y1−3y2+y3−4y1+4y2+10y32y1+2y2−4y3=10,=20,=43.
Solving this 3×3 system:
y1=501011=20.22,y2=50596=25298=11.92,y3=50266=25133=5.32.
All yi>0, so this is a feasible BFS, and one checks the simplex optimality test (all reduced costs ≤0 for the maximization) holds here.
Step 5 — Back-substitute to original variables
x1=y1+2=501011+2=501111=22.22,x2=y2+1=25298+1=25323=12.92,x3=y3+3=25133+3=25208=8.32.
Optimal objective:
z=−6x1−2x2−5x3=−6⋅501111−2⋅25323−5⋅25208=−255019=−200.76.
Answer
x1=501111=22.22, x2=25323=12.92, x3=25208=8.32,zmin=−255019=−200.76.