- [[Greedy Algorithms]]
- [[TSP OPT Approximation Algorithm]]
- [[Vertex Cover Approximation Algorithm]]
- in optimization problems we are asked to find a solution with the maximum gain or the minimum cost
- denote this optimum value (benefit or cost) by $OPT(I)$
- falls short of NP complete problem optimum, but never by too much
- denote the approximation value on input $I$ as $A(I)$
- these algorithms are used to get near optimal solutions but in polynomial time
- want to bound how far away from optimal it can get in worst case
- the approximation ratio is then:
- $\alpha = \frac{A{(I)}}{OPT{(I)}}$
- for the worst case scenario input $I$, the input that makes the approximate solution as far away from the actual optimum as possible
- for a maximizing problem, ratio will be a fraction, goal is for $A(I)$ to reach the bigger bottom
- for a minimizing problem, ratio will be a number
gt;1$, goal is for $A(I)$ to reduce to the smaller bottom
- in both cases the goal is to get a ratio of 1 for the value
https://www.geeksforgeeks.org/approximation-algorithms/#
https://people.eecs.berkeley.edu/~vazirani/algorithms/chap9.pdf
https://www.youtube.com/watch?v=tutKR8AV23c&list=PLkFD6_40KJIyKLUW_4cm44mIdXSUpZry3&index=21