Greedy-HPS-MOCO v1.0.0
This is the initial release of the Greedy Hypervolume Polychotomic Scheme for Multiobjective Combinatorial Optimization (Greedy-HPS-MOCO).
About
The Greedy-HPS-MOCO is mainly designed to solve the Multiobjective Knapsack Problem (MOKP), maximizing the hypervolume dominated by a subset of nondominated solutions of size J with respect to K reference points.
There are three solver variants available:
- Greedy: The main solver that uses a greedy algorithm to find a representation set. In this case, the non-dominated set is required to be known.
- ILP: An alternative solver that uses a mixed-integer programming formulation for hypervolume scalarization.
- BB: A specialized branch-and-bound solver for m-objective knapsack problems.
In both the ILP and BB solvers, the non-dominated set is not required to be known, and the algorithm will compute it during the solution process. Moreover, for both solvers, it is necessary to provide the number of reference points K to be used in the hypervolume scalarization.
For detailed information about changes and improvements in this release, please see the UPDATE.md file.