Published May 8, 2020 | Version v1

Supplementary Data for Multiform: Multi-objective Evolution of Requirements Models Constrained by Formal Verification Results

Authors/Creators

Description

This data set provides supplementary material for the article "Multiform: Multi-objective Evolution of Requirements Models Constrained by Formal Verification Results" (to appear). It contains the following files:

  • experiment-input-models.zip which contains the SML input models for the EBEAS and production cell examples that were used to conduct the experiment
  • experiment-results.tar which contains the computed candidate SML models as well as H2 database files that contain measurements.
  • experiment-results.pdf which summarizes the conducted controlled experiment and results.

 

Input models (example for EBEAS)

  • ebeas.sml contains the actual SML input model to be evolved
  • ebeas.ecore contains the metamodel of the EBEAS example
  • ebeas.xmi contains the object system of the EBEAS example that is used for the SML realizability check
  • ebeas.runconfig contains the runtime configuration for ScenarioTools that binds the SML input model with the object system
  • ebeas.cspec contains the solution space model used by Multiform.

 

Measurements are stored in an H2 database file. To open one of the database files for the EBEAS or production cell examples extract the appropriate zip file to a local folder, navigate to the folder in a terminal, and start H2 with the appropriate database file as follows:

java -jar h2-1.4.199.jar -url jdbc:h2:./Statistics

A web-based SQL client will open in your browser. H2 cann be obtained free of charge from their homepage.

 

The database schema consists of three simple tables:

SMLCANDIDATESTATISTICS contains measurements for each evolved candidate SML model and consists of the following columns:
ALGORITHM - one of 'Random', 'NSGA2', 'tabu-75-intensify'
SEED - seed id for which the measurement was taken
ITERATION - iteration id during whch the measurement was taken
CANDIDATE - unique id of the evaludated candidate SML model
SYNTHESISTIME - synthesis time of the evaludated candidate SML model
O1_SCENARIOS - objective value for o1
O2_FRAGMENTSRATIO - objective value for o2
O3_ENVFRAGMENTSRATIO - objective value for o3
C1_REALIZABILITY - constraint value for c1
C2_REACHABILITY - constraint value for c1

SMLITERATIONSTATISTICS contains aggregated statistical data for each iteration and consists of the following columns:
ALGORITHM - one of 'Random', 'NSGA2', 'tabu-75-intensify'
SEED - seed id for which the data was aggregated
ITERATION - unique id of this aggregated iteration data
ITERATIONSUCCESSRATE - achieved success rate in this iteration
ACCUMULATEDSUCCESSRATE - achieved aggregated success reate until this iteration
ACCUMULATEDHYPERVOLUMEINDICATOR - achieved hypervolume until this iteration
NUMITERATIONPARETOEQUIVALENTCANDIDATES - number of pareto-equivalent candidate SML models in this iteration
NUMACCUMULATEDPARETOEQUIVALENTCANDIDATES - number of pareto-equivalent candidate SML models until this iteration
NUMITERATIONPARETODOMINANTCANDIDATES - number of pareto-dominant candidate SML models in this iteration
NUMACCUMULATEDPARETODOMINANTCANDIDATES - number of pareto-dominant candidate SML models until this iteration
ITERATIONSYNTHESISTIME - total synthesis time of this iteration
ACCUMULATEDSYNTHESISTIME - accumulated total synthesis time until this iteration

SMLSEEDSTATISTICS contains aggregated statistical data for each seed and consists of the following columns:
ALGORITHM - one of 'Random', 'NSGA2', 'tabu-75-intensify'
SEED - unique id of this aggregated seed data
SUCCESSRATE- achieved success rate in this seed
HYPERVOLUMEINDICATOR - achieved hypervolume in this seed
NUMPARETOEQUIVALENTCANDIDATES - number of pareto-equivalent candidate SML models in this seed
NUMPARETODOMINANTCANDIDATES - number of pareto-dominant candidate SML models in this seed
TOTALSYNTHESISTIME - total synthesis time of this seed

 

Please note: the database files contain data for algorithms 'tabu-50-intensify' and 'tabu-25-intensify' representing evaluation runs with different Tabu search configurations. However, these still need to be analyzed and experiment-results.pdf refers to 'tabu-75-intensify' only.

Files

experiment-input-models.zip

Files (2.4 GB)

Name Size
md5:9c60fdb7e866ac231234f44621335539
9.4 kB Preview Download
md5:0fb89e0b8b98a3c46ac3ecc3df85d1bc
1.6 MB Preview Download
md5:0a43501a5a3259f01cc266f3c5eac588
2.4 GB Download