Shifting lemma: the best of shifted optima is a -approximation
ProvedWilliamsonShmoys.shifting_boundapproximation-algorithmsindependent-setplanar-graphs
Let be a simple graph on a vertex set , let be arbitrary real weights, let , and let be any labelling (in Baker's technique, the breadth-first-search level). For write . Suppose a finite set has weight at least that of every independent set avoiding some class, i.e.
Then for every finite independent set ,
This is the averaging argument behind Baker's shifting technique in the proof of Theorem 10.11: the sets are independent and together cover every vertex of exactly times, so their total weight is and one of them has weight at least .
Preamble
import Mathlib.Combinatorics.SimpleGraph.Clique import Mathlib.Data.Real.Basic import Mathlib.Algebra.BigOperators.Group.Finset.Basic set_option autoImplicit false open scoped BigOperators
Formal statement
namespace WilliamsonShmoys
theorem shifting_bound {V : Type} (G : SimpleGraph V) (w : V → ℝ) (c : V → ℕ)
(k : ℕ) (hk : 0 < k) (A : Finset V)
(hA : ∀ j < k, ∀ S : Finset V, G.IsIndepSet (S : Set V) →
(∀ v ∈ S, c v % k ≠ j) → ∑ i ∈ S, w i ≤ ∑ i ∈ A, w i)
(S : Finset V) (hS : G.IsIndepSet (S : Set V)) :
(1 - 1 / (k : ℝ)) * ∑ i ∈ S, w i ≤ ∑ i ∈ A, w i := by sorry
end WilliamsonShmoys
Source
David P. Williamson and David B. Shmoys, The Design of Approximation Algorithms, Cambridge University Press, 2011, author electronic manuscript, Section 10.2, proof of Theorem 10.11, pp. 270-272. https://doi.org/10.1017/CBO9780511921735. (the averaging step: some shift loses at most a 1/k fraction of the optimum)