click below
click below
Normal Size Small Size show me how
PDS #4
| Question | Answer |
|---|---|
| What is the notation and decision variables in Matching Problems | c(i,j) ...cost (perfect match) or profit (matching) of employee i and j working together |
| What are decision variables in Matching Problems | x(i,j) ∈ {1 i,j working together 0 otherwise} |
| What are conditions in Matching and Perfect Match | Matching >> ∑x(i,j) <=1 Perfect Match >> ∑x(i,j) = 1 |
| What are 3 types of Set Problems | 1, Set partitioning = every element must be contained exactly once (b(i,j)=1) 2, Set packing = every element can be contained at most once (b(i,j)<=1) 3, Set covering = every element must be contained at least once (b(i,j)>=1) |
| What are objective of the 3 types of set problems | Set Partitioning >> cost minimization (and have every element exactly once) Set Packing >> profit maximization (and having every element at most once) Set Covering >> cost minimization (and having each element at least once) |
| What are notation for Set Problems(4) | 1, M = {1,..n} set M with all the elements 2, M(j) ,, subset j containing some of the elements 3,c(j) = cost/profit of subset j 4, b(i,j) = 1 or 0 if element i s in subset j |
| What is generalized incidence matrix for Set Problems | Subset are rows M1, M2, ... Mn elements are columns 1 2 3 n 1 if element inside of subset / 0 otherwise |
| What is notation for Facility Location Problem (4) | 1, I∈{1,2, ..n} set of all costumers 2, J∈{1,2...n} set of all facilities 3, c(i,j) .. cost of using facility j to serve costumer i 4,f(j) .. cost of opening facility j |
| What are 2 decision variables for Facility Location Problem | 1, y(j) ∈{0,1} .. should we open facility j 2, x(i,j) ∈{0,1≠ .. should we serve i from j |
| What are constrains for Facility Location Problem | x(i,j) <= y(j) ... from facility j can be served only if its open |
| What is objective function for Facilitiy Location Problem | z* = min ∑f(j)y(j) + ∑∑ c(i,j)x(i,j) |
| How to write matematically for facility to be opened it must serve at lease two costumers | ∑x(i,j)>= 2y(j) ∀j∈J |
| What is LP (Linear Programming) -Relaxation of MIP (Multiple Integer Programming) (3) | 1, change Integral Problem into Linear Problem with x^n ∈ R 2, also binary variable can se set as float between 0 and 1 3, LP Relaxation give us upper bound for optimal solution |
| Can we use LP Relaxation for solution | Not directly, NO ROUNDING UP/DOWN |
| What is convex hull (3) | 1, X = conv(s) 2, all extreme points of X =conv(S) are integer 3, complete description is NOT KNOWN in general |
| What is the idea of Cutting Plane Method | adding valid inequalitites so that the feasible region converge toward convex hull |
| What is the sequence for cutting plane method (5) | 1, Solve the LP-Relaxation with x̂ as solution 2, Check if x̂ has only integer values yes > x̂ is optimal / no > step 3 3, Determine a violated valid inequality 4, add violated valid inequality as constrain 5,repeat from step 2 |
| What do we do in step 3 in cutting plane method (2) | 1, a^⊤ x <= b ∀x∈S but a^⊤x̂ >b 2, cutting as much as possible from feasible region but not cutting out any integer solution |