4. Mixed integer problems Flashcards
Fig. definition MIP
- Z(X) denotes the optimal objective value over the feasible set X
- x and y denote the n-dimensional vector of non-negative continuous variables and the p-dimensional vector of non-negative integer variables, respectively
Fig. Set of feasible solutions MIP
- c ∈ Rn and f ∈ Rp are the row vectors of objective coefficients
- b ∈ Rm is the column vector of right-hand-side coefficients of the m constraints
- A and B are the matrices of constraints with real coefficients of dimensions (m × n) and (m × p), respectively
Fig. definition Mixed Binary Linear Program (MBP)
Linear Relaxation (LR) of a Mixed Integer Program
obtained by removing the integrality restrictions on the y variables
Key points LR
1: If in the solution to the linear relaxation all variables (that should be integer) have integer values, then it is also an optimal solution for the (mixed) integer problem
2: The optimal objective value for the linear relaxation will be lower than the one of the (mixed) integer program
Lower bounds for a MIP
minimization -> linear relaxation
maximization -> any feasible solution
Upper bounds for a MIP
minimization -> any feasible solution
maximization -> linear relaxation
Fig. X and Px for MIP
Running time of an algorithm
- defined as the number of elementary operations performed
- elementary operations are the elementary arithmetic operations (+, −, ×, /) and the comparison of two numbers.