A bound on the shannon capacity via a linear programming variation

We prove an upper bound on the Shannon capacity of a graph via a linear programming variation. We also show that our bound can be better than Lovasz theta number and Haemers minimum rank bound.