Two Phase Method Flashcards
phase 1 of two phase method
- add artificial variables to constraints to force an identity matrix (immediate BFS in the form x=0, xₐ=b)
- solve the following LP problem
min eᵀxₐ s.t. Ax + xₐ = b x,xₐ≥0 - if at optimality xₐ is not equal to zero then the problem has no feasible solution. Otherwise the optimal soltuion for the phase 1 problem is (x,0) so x is feasible for the original.
phase 2 of two phase method
- x* is feasible for the original problem. Split x* into basic and non basic.
- remove the artificial variables from the table
- restore the simplex tableu by replacing the zero row with zⱼ-cⱼ
- carry out the simplex algorithm
Big M method
- introduce artificial variables to force an identity matrix
- solve the following LP problem
min cᵀx + M(eᵀxₐ) s.t. Ax + xₐ = b x,xₐ≥0
-> if an optimal solution is found and xₐ=0 then x is the optimal solution to the original problem
-> if an optimal solution is found and xₐ/= 0. The the original problem is infeasible
-> if an optimal solution is unbounded and xₐ=0 then the original problem is unbounded
-> if an optimal solution is unbounded amd x*ₐ/=0 then the original problem is unbounded
if xₐ/=0 in two phase and big M
the original problem is unbounded
Big M: if xₐ=0 and there is a optimal solution
x* is the optimal solution of the original LP problem
Big M: if xₐ=0 and the optimal solution is unbounded
the original problem is also unbounded
Two Phase: if xₐ=0 and all the artificial variables are out of the basis
remove xₐs, restore the simplex tableu and carry out simplex
Two phase: if xₐ=0 and some of the artificial variables are still in the basis
just remove the row corresponding to the artificial variable (if its value in the RHS column if zero)
How can you tell if xₐ=0
xₐ=0 if the corresponding value in the RHS is 0
if xₐ not in the basis then it automatically has xₐ=0