How to calculate the shadow price in Gurobi

问题

I want to analyze whether the boundary should increase or reduce in Constraints in a programming problem: The following is simplified problem. V[(i,t)]is decision variable and S[i] is input. I want to know if the obj increases or reduces when increasing one unit of S[i]`. I know may the shadow price and marginal cost are for decision variable not inputs. In Gurobi, Dual value (also known as the shadow price) can use the Pi function.

for t in range(T):
    for i in range(I):
        m.addConstr(V[(i,t)] <= Lambda*S[i])
        m.addConstr(other constrints without S[i])
obj =cf*quicksum(V[(i,0)] for i in range(I))+ cs*quicksum(S[i]for i in range(I))+...
m.setObjective(obj, GRB.MAXIMIZE)
m.optimize()

回答1:

There are two ways to get the shadow price:(Python + Gurobi):

shadow_price = model.getAttr('Pi', model.getConstrs())

or

shadow_price = model.getAttr(GRB.Attr.Pi)

It returns the shadow prices of all constraints in sequence into an array.

来源：https://stackoverflow.com/questions/49126492/how-to-calculate-the-shadow-price-in-gurobi