Skip to main content

Using dictionaries as objective coefficients

Answered
Binghao Zhang

I'm fairly new to Gurobi and have been going over the available examples but I'm unsure how to adapt the examples to my specific problem. I have several (100+) dictionaries which I would like to turn into an objective function. As a small toy example: 

d = {(111,222,333):20, (111,222,444):30, (111,222,555):10, (111,222,666):0}
k1 = {(111,222,333):0.5, (111,222,444):0.1, (111,222,666):0.25, (111,222,555):0.3}
k2 = {(111,222,333):0.1, (111,222,555):0.3, (111,222,666):0.1, (111,222,444):0.5}
k3 = {(111,222,444):0.4, (111,222,555):0, (111,222,666):0.2, (111,222,333):0.6}

All of the dictionaries are the same size (around 20000 keys), with the same keys. I'd like to formulate an absolute value problem with the dictionaries as the objective coefficients. For the toy example, it would be something like:

where D, K1, K2, K3 refer to the respective dictionaries, i refers to the keys and x are the decision variables. Is there a way to easily do this? Would really appreciate any help or if anyone could point me towards the right direction.

 

 

0

Comments

2 comments

Date Votes
  • Official comment
    Simranjit Kaur
    • Gurobi Staff
    This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum, or try Gurobot, our chatbot interface offering instant, expert-level support.
  • Matthias Miltenberger
    • Gurobi Staff
    Edited

    Hi Bing,

    This should do the trick:

    import gurobipy as gp

    m = gp.Model()

    d = {(111,222,333):20, (111,222,444):30, (111,222,555):10, (111,222,666):0}
    k1 = {(111,222,333):0.5, (111,222,444):0.1, (111,222,666):0.25, (111,222,555):0.3}
    k2 = {(111,222,333):0.1, (111,222,555):0.3, (111,222,666):0.1, (111,222,444):0.5}
    k3 = {(111,222,444):0.4, (111,222,555):0, (111,222,666):0.2, (111,222,333):0.6}

    x_d = m.addVars(d, name="d")
    x_d_abs = m.addVars(d, name="d_abs")
    x_k1 = m.addVars(k1, name="k1")
    x_k2 = m.addVars(k2, name="k2")
    x_k3 = m.addVars(k2, name="k3")

    for i in d.keys():
        m.addConstr(x_d[i] == d[i] - k1[i]*x_k1[i] - k2[i]*x_k2[i] - k3[i]*x_k3[i])
        m.addGenConstrAbs(x_d_abs[i], x_d[i])

    m.setObjective(gp.quicksum(x_d_abs))
    m.write("dict.lp")

    You just have to define a set of auxiliary variables to hold the absolute values for the objective.

    I added a call to write() so you can inspect the final model to see whether this is actually correct.

    Cheers,
    Matthias

    0

Post is closed for comments.

Didn't find what you were looking for?

New post