Metaheuristic-based Extensible Tool for Cooperative Optimisation (METCO) -- Version registered at Safe Creative
Currently, METCO only allows two formats for the results files, plain text and XML. The differences between each format and how to interpret them are explained down below.
Let's supposed that we are trying to find the global optimum of the Ackley Function (30 dimensions instance) using the Differential Evolution algorithm implemented in METCO. Notice that the Ackley Function it is not implemented in the latest version of METCO. We should run our experiment like:
./metcoSeq /home/username . PlainText results.txt MonoDE Ackley EVALUATIONS 100000 1000 0 0.5 0.115 30 ! 30 $ NoOp;
After it finishes, we will have the result.txt file with the results of the experiment for every 1000 evaluations of the algorithm. If you don't realize why it is this way, check out how to run METCO.
Next, there is an example of the result file
Current Evaluation = 1000(1000)
Front Size = 1
4.14602569773142449 10.1111731614466649 -7.0645145754332308 5.28332344110465524 -16.4433158111848527 -19.38260894152085 -1.53641686320325022 3.04215541981260351 -32.7680000000000007 -14.8349038612781605 -28.78761358457227 16.8081287233420369 32.7680000000000007 -3.09718953467588065 -24.5689186668363035 -19.1237361386659259 20.0588857362693602 -5.64246610141288318 -32.6682047118675953 -6.03445241827631662 -32.7680000000000007 -32.7680000000000007 32.7680000000000007 -11.7832334497008624 -28.6204749126194393 -32.7680000000000007 -20.1861851748052032 6.89767668473053419 -0.531230880984679743 17.1323408588079502 21.7182817872359735
Current Evaluation = 2000(2000)
Front Size = 1
-24.35348208226333 -32.7680000000000007 -8.50852747168975654 -7.35352755418119219 -32.7680000000000007 -24.6259077116483347 27.7638852836433259 27.6237015356904365 -1.60598143843095187 -7.13597008083759121 -24.6687523270085229 27.7180705161420953 -32.7680000000000007 7.72639757134690619 25.9290196501356647 32.7680000000000007 32.7680000000000007 6.15664724636666882 19.4532477660166414 -3.32440732729268262 -29.3155716412680079 32.7680000000000007 -12.3974223137411332 -6.06670901008821595 32.4274260862671895 15.1634840921440635 -13.7380409152107479 -32.0627234226757309 28.9572169925110501 -32.7680000000000007 21.7182817872359735
(...)
Genetic Algorithm DE
Number of Evaluations = 100000
Weight factor = 0.5
Crossover = 0.100000001490116119
Population Size = 30
Time = 0.468252999999999975
Current Evaluation = 100000
Front Size = 1
32.7680000000000007 -15.9025481981702868 32.7680000000000007 -16.0435141713351683 -11.9267453539161483 -8.90274492002194684 -19.0898707743117981 20.1607823281172145 9.15719495930455274 -28.770813669048227 10.4724050928404644 -30.2265264212207967 0 -19.3500226426485327 -16.6668501665244193 -18.9441218480358593 5.213448917308666 -1.21794220431338829 20.7714501836740055 10.2722986055862968 -19.4581973897388814 -11.5055220812062799 1.07924994624690029 24.6890939986446298 6.61085563166696843 -19.7870624185111659 1.59072514066614978 32.7680000000000007 -25.0643355725077868 -18.8879885874006135 21.7182817872359735
Notice that the results are not sorted in any particular way.
The result files follows this format:
When the print period is reached, METCO shows the status of the execution:
- Current Evaluation
- Number of items of the Pareto's Front. In this case, as we are working with a single-objective function, the value is one.
- MOFront returned from the
getSolution(MOFront*)method inside the algorithm class. In this case DE returns the evaluation of every single individual of the population.
This information is printed continuosly after the algorithm performs printPeriod evaluations.
In the end, when the execution is finished; besides the finals results, METCO shows the elapsed time in this execution and the algorithm information described in the method printInfo of the algorithm.
Coming soon...