SPA: series SPA2021

There are currently 49 data files.

Format

The format is originally from the OR library extended for handling multiple objectives.

The format of all of these data files is:

number of rows (m), number of columns (n)   
for each column j (j=1,...,n) in turn:    
    column cost for objective 1 c1(j)
    column cost for objective 2 c2(j)
    number of rows covered by j
    list of the rows covered by j

Instances:

The data files are:

• biodidactic
• biosppaa01, biosppaa02, biosppaa03, biosppaa05
• biosppnw01, biosppnw02, … , biosppnw43
• biorail4284

For these instances, the coefficients of the second objective are a random permutation of the coefficients of the first objective.

• Click to access these files

The set of non dominated points have been computed with vOptGeneric for several instances.

• Click to access these files

The format of all of these files is:

CPU time   
number of non-dominated points (K)
for each non-dominated point (k=1,...,K) in turn:
    value on the 1st objective, z1(k)
    value on the 2nd objective, z2(k)

Parser:

A parser is available in the folder “examples” of vOptGeneric (function loadInstance2SPA in the file setPartitionning.jl).

---

**Instances referenced in: **

Xavier Gandibleux, Guillaume Gasnier and Saïd Hanafi. A primal heuristic to compute an upper bound set for multi-objective 0-1 linear optimisation problems. MODeM ‘21: 1st Multi-Objective Decision Making Workshop, July 14-16, 2021. http://modem2021.cs.nuigalway.ie/