set3-20: Instance-to-Instance Comparison Results

Type: Instance
Submitter: Kerem Akartunali
Description: Multi-item lot-sizing with backlogging. Solved by SCIP 3.1.1 parallelized by UG 0.7.5 linked to CPLEX 12.6 as an LP solver on HLRN III (https://www.hlrn.de/home/view/System3/CrayHardware). Due to time limit restrictions, four repeated runs, each starting from the checkpointing file of the previous run, were done. Each run had a time limit of 12 hours while using 6144 (runs 1) or 3072 cores (run 2 - 4).
MIPLIB Entry

Parent Instance (set3-20)

All other instances below were be compared against this "query" instance.

set3-20 Raw set3-20 Decomposed set3-20 Composite of MIC top 5 set3-20 Composite of MIPLIB top 5 set3-20 Model Group Composite
Raw This is the CCM image before the decomposition procedure has been applied.
Decomposed This is the CCM image after a decomposition procedure has been applied. This is the image used by the MIC's image-based comparisons for this query instance.
Composite of MIC Top 5 Composite of the five decomposed CCM images from the MIC Top 5.
Composite of MIPLIB Top 5 Composite of the five decomposed CCM images from the MIPLIB Top 5.
Model Group Composite Image Composite of the decomposed CCM images for every instance in the same model group as this query.

MIC Top 5 Instances

These are the 5 decomposed CCM images that are most similar to decomposed CCM image for the the query instance, according to the ISS metric.

Decomposed These decomposed images were created by GCG.
sing5 decomposed neos-3072252-nete decomposed control20-5-10-5 decomposed neos2 decomposed sing11 decomposed
Name sing5 [MIPLIB] neos-3072252-nete [MIPLIB] control20-5-10-5 [MIPLIB] neos2 [MIPLIB] sing11 [MIPLIB]
Rank / ISS The image-based structural similarity (ISS) metric measures the Euclidean distance between the image-based feature vectors for the query instance and all other instances. A smaller ISS value indicates greater similarity.
1 / 1.049 2 / 1.056 3 / 1.058 4 / 1.060 5 / 1.070
Raw These images represent the CCM images in their raw forms (before any decomposition was applied) for the MIC top 5.
sing5 raw neos-3072252-nete raw control20-5-10-5 raw neos2 raw sing11 raw

MIPLIB Top 5 Instances

These are the 5 instances that are most closely related to the query instance, according to the instance statistic-based similarity measure employed by MIPLIB 2017

Decomposed These decomposed images were created by GCG.
neos-5041822-cockle decomposed set3-16 decomposed set3-09 decomposed set3-15 decomposed set3-10 decomposed
Name neos-5041822-cockle [MIPLIB] set3-16* [MIPLIB] set3-09* [MIPLIB] set3-15* [MIPLIB] set3-10* [MIPLIB]
Rank / ISS The image-based structural similarity (ISS) metric measures the Euclidean distance between the image-based feature vectors for the query instance and all model groups. A smaller ISS value indicates greater similarity.
650 / 1.735 0* / 0.000* 0* / 0.000* 0* / 0.000* 0* / 0.000*
Raw These images represent the CCM images in their raw forms (before any decomposition was applied) for the MIPLIB top 5.
neos-5041822-cockle raw set3-16 raw set3-09 raw set3-15 raw set3-10 raw

Instance Summary

The table below contains summary information for set3-20, the five most similar instances to set3-20 according to the MIC, and the five most similar instances to set3-20 according to MIPLIB 2017.

INSTANCE SUBMITTER DESCRIPTION ISS RANK
Parent Instance set3-20 [MIPLIB] Kerem Akartunali Multi-item lot-sizing with backlogging. Solved by SCIP 3.1.1 parallelized by UG 0.7.5 linked to CPLEX 12.6 as an LP solver on HLRN III (https://www.hlrn.de/home/view/System3/CrayHardware). Due to time limit restrictions, four repeated runs, each starting from the checkpointing file of the previous run, were done. Each run had a time limit of 12 hours while using 6144 (runs 1) or 3072 cores (run 2 - 4). 0.000000 -
MIC Top 5 sing5 [MIPLIB] Daniel Espinoza Imported from the MIPLIB2010 submissions. 1.048635 1
neos-3072252-nete [MIPLIB] Jeff Linderoth (None provided) 1.056257 2
control20-5-10-5 [MIPLIB] Qie He Optimal control of a discrete-time switched system model Numerically challenging. Different solvers report this instance as solved to optimality, infeasible, or unbounded. 1.058048 3
neos2 [MIPLIB] NEOS Server Submission Imported from the MIPLIB2010 submissions. 1.059596 4
sing11 [MIPLIB] Daniel Espinoza Imported from the MIPLIB2010 submissions. 1.070085 5
MIPLIB Top 5 neos-5041822-cockle [MIPLIB] Jeff Linderoth (None provided) 1.734889 650
set3-16* [MIPLIB] Kerem Akartunali This is joint work with Andrew J. Miller, Universite de Bordeaux 1; RealOpt, INRIA Bordeaux Sud-Ouest, France. Andrew.Miller@math.u-bordeaux1.fr. These are five hard instances chosen from our library MULTILSB (Multi-item Lot-Sizing with Backlogging), which is fully available at (incl. documentation and mosel file):http://personal.strath.ac.uk/kerem.akartunali/research/multi-lsb/ For these 5 instances, gaps remain in the range of >50% after 30 min. (using latest version of FICO on a laptop with i7 processor and 4GB RAM): (first two columns indicate the 30 min-runs, and the last column indicates the best solutions we know at the moment). Computational results for these instances as well as for others of MULTILSB can be found on the website. 0.000000* 0*
set3-09* [MIPLIB] Kerem Akartunali This is joint work with Andrew J. Miller, Universite de Bordeaux 1; RealOpt, INRIA Bordeaux Sud-Ouest, France. Andrew.Miller@math.u-bordeaux1.fr. These are five hard instances chosen from our library MULTILSB (Multi-item Lot-Sizing with Backlogging), which is fully available at (incl. documentation and mosel file):http://personal.strath.ac.uk/kerem.akartunali/research/multi-lsb/ For these 5 instances, gaps remain in the range of >50% after 30 min. (using latest version of FICO on a laptop with i7 processor and 4GB RAM): (first two columns indicate the 30 min-runs, and the last column indicates the best solutions we know at the moment). Computational results for these instances as well as for others of MULTILSB can be found on the website. 0.000000* 0*
set3-15* [MIPLIB] Kerem Akartunali Multi-item lot-sizing with backlogging. The problem was solved with CPLEX 12.4 in approximately 64.5 hours. 0.000000* 0*
set3-10* [MIPLIB] Kerem Akartunali Multi-item lot-sizing with backlogging. Solved by SCIP 3.1.1 parallelized by UG 0.7.5 linked to CPLEX 12.6 as an LP solver on HLRN III (https://www.hlrn.de/home/view/System3/CrayHardware). Due to time limit restrictions, five repeated runs, each starting from the checkpointing file of the previous run, were done. Each run had a time limit of 12 hours while using 6144 (runs 1, 3, and 4) or 3072 cores (run 2 and 5). 0.000000* 0*


set3-20: Instance-to-Model Comparison Results

Model Group Assignment from MIPLIB: set3
Assigned Model Group Rank/ISS in the MIC: 1 / 0.0

MIC Top 5 Model Groups

These are the 5 model group composite (MGC) images that are most similar to the decomposed CCM image for the query instance, according to the ISS metric.

These are model group composite (MGC) images for the MIC top 5 model groups.
Model group: set3 Model group: neos-pseudoapplication-2 Model group: sp_product Model group: noip Model group: allcolor
Name set3 neos-pseudoapplication-2 sp_product noip allcolor
Rank / ISS The image-based structural similarity (ISS) metric measures the Euclidean distance between the image-based feature vectors for the query instance and all other instances. A smaller ISS value indicates greater similarity.
1 / 0.000 2 / 1.496 3 / 1.499 4 / 1.556 5 / 1.582

Model Group Summary

The table below contains summary information for the five most similar model groups to set3-20 according to the MIC.

MODEL GROUP SUBMITTER DESCRIPTION ISS RANK
MIC Top 5 set3 Kerem Akartunali Multi-item lot-sizing with backlogging. Solved by SCIP 3.1.1 parallelized by UG 0.7.5 linked to CPLEX 12.6 as an LP solver on HLRN III (https://www.hlrn.de/home/view/System3/CrayHardware). Due to time limit restrictions, four repeated runs, each starting from the checkpointing file of the previous run, were done. Each run had a time limit of 12 hours while using 6144 (runs 1) or 3072 cores (run 2 - 4). 0.000000 1
neos-pseudoapplication-2 NEOS Server Submission Imported from the MIPLIB2010 submissions. 1.495529 2
sp_product MIPLIB submission pool Imported from the MIPLIB2010 submissions. 1.498902 3
noip Christopher Hojny integer programming formulation that verifies that no integer programming formulation of a given 0/1-point set exists 1.555506 4
allcolor Domenico Salvagnin Prepack optimization model. 1.582203 5


* set3-16 is a duplicate of set3-20. set3-09 is a duplicate of set3-20. set3-15 is a duplicate of set3-20. set3-10 is a duplicate of set3-20.

Updated August 16, 2019