I have a set partitioning formulation with each constraint being an equality constraint to meet the given demand (right-hand side of the constraint). For each constraint, I have a slack and a surplus variables to penalize under- and over-coverage in the objective function. The presence of slack and surplus variables makes initial basis readily available to start simplex algorithm.

When solving this formulation with coin-or lp solver CLP, I would like to indicate this initial basis. I tried to set ClpSimplex::setColumnStatus(), but the subsequent call of ClpSimplex::primal() seems to ignore such a suggestion, introduces an auxiliary variable for each constraint, and creates additional degeneracy. Moreover, for some constraints with both slack and surplus variables equal to zero in the found optimal solution, the corresponding auxiliary variable is still part of the found optimal basis.

Is there any way to manually indicate initial basis in CLP so that the solver does not introduce auxiliary variables?

I have not tried other solvers like CPLEX or Gurobi. Do they provide the possibility to manually indicate initial basis?


1 Answer 1


in CPLEX you can use warmstart with all APIs. Let me show you in OPL:


int nbKids=300;

// a tuple is like a struct in C, a class in C++ or a record in Pascal
tuple bus
 key int nbSeats;
 float cost;

// This is a tuple set
{bus} pricebuses=...;

// asserts help make sure data is fine
assert forall(b in pricebuses) b.nbSeats>0;assert forall(b in pricebuses) b.cost>0;

// To compute the average cost per kid of each bus
// you may use OPL modeling language

float averageCost[b in pricebuses]=b.cost/b.nbSeats;

// Let us try first with a naïve computation, use the cheapest bus

float cheapestCostPerKid=min(b in pricebuses) averageCost[b];
int cheapestBusSize=first({b.nbSeats | b in pricebuses : averageCost[b]==cheapestCostPerKid});
int nbBusNeeded=ftoi(ceil(nbKids/cheapestBusSize));

float cost0=item(pricebuses,<cheapestBusSize>).cost*nbBusNeeded;
execute DISPLAY_Before_SOLVE
  writeln("The naïve cost is ",cost0);
  writeln(nbBusNeeded," buses ",cheapestBusSize, " seats");

int naiveSolution[b in pricebuses]=

// decision variable array
dvar int+ nbBus[pricebuses];

// objective
  sum(b in pricebuses) b.cost*nbBus[b];

// constraints
subject to
  sum(b in pricebuses) b.nbSeats*nbBus[b]>=nbKids;

float cost=sum(b in pricebuses) b.cost*nbBus[b];
execute DISPLAY_After_SOLVE
 writeln("The minimum cost is ",cost);
 for(var b in pricebuses) writeln(nbBus[b]," buses ",b.nbSeats, " seats");


// Warm start the naïve solution



and then we get

The naïve cost is 4000
8 buses 40 seats

The minimum cost is 3800
6 buses 40 seats
2 buses 30 seats

and in the cplex log we see

1 of 1 MIP starts provided solutions.
MIP start 'm1' defined initial solution with objective 4000.0000.

The warmstart is in the line

// Warm start the naïve solution

and many other similar tiny examples at https://www.linkedin.com/pulse/making-decision-optimization-simple-alex-fleischer/

  • $\begingroup$ Thank you, Alex! $\endgroup$
    – Sergei
    Apr 29, 2020 at 5:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.