ttc18 issueshttps://git-st.inf.tu-dresden.de/stgroup/ttc18/-/issues2018-06-26T20:44:34+02:00https://git-st.inf.tu-dresden.de/stgroup/ttc18/-/issues/3EMFeR: Testcases do not pass2018-06-26T20:44:34+02:00René SchöneEMFeR: Testcases do not passRunning `:jastadd-mquat-solver-emfer:test`, the five test cases of `EMFeRHandwrittenTest` fail.
The solver always returns an empty solution (no assignments), thus the assertions fail.
Are there any assumptions in the solver w.r.t. the s...Running `:jastadd-mquat-solver-emfer:test`, the five test cases of `EMFeRHandwrittenTest` fail.
The solver always returns an empty solution (no assignments), thus the assertions fail.
Are there any assumptions in the solver w.r.t. the structure of the model. They may not hold for the testcases.https://git-st.inf.tu-dresden.de/stgroup/ttc18/-/issues/2Review of ACO and ILP by Albert Zündorf2018-06-26T20:44:48+02:00Albert ZündorfReview of ACO and ILP by Albert ZündorfT read both papers. For execution I rely on the results of the benchmark execution already given.
I like the ACO solution very much. I am not really sure whether I got all the details but
the idea of having 'pheromones' guiding the ex...T read both papers. For execution I rely on the results of the benchmark execution already given.
I like the ACO solution very much. I am not really sure whether I got all the details but
the idea of having 'pheromones' guiding the exploration towards promising areas seems to be
quite appropriate for the given case.
The ILP solution ofcourse works great, too.
However, just from my feelings the ACO approach seems
to be the better fit for the given problem.https://git-st.inf.tu-dresden.de/stgroup/ttc18/-/issues/1Review: ACO, EMFeR and ILP Solvers2018-06-26T20:45:00+02:00Samaneh HoseinDoostReview: ACO, EMFeR and ILP SolversHello,
I run ACO, EMFeR and ILP solvers using GLPK, on a standard Windows 7 PC using an Intel®Core™ i5-2430M with 2.40GHz processor and 4.00 GB RAM.
The experimental results for five benchmark were attached here:
[Exprimental_Results.p...Hello,
I run ACO, EMFeR and ILP solvers using GLPK, on a standard Windows 7 PC using an Intel®Core™ i5-2430M with 2.40GHz processor and 4.00 GB RAM.
The experimental results for five benchmark were attached here:
[Exprimental_Results.pdf](/uploads/d816636112a5a4720e412e814b2cbb83/Exprimental_Results.pdf)
Also, the evaluation sheets and solution models are available here:
[https://github.com/Ariyanic/TTC18/tree/master/jastadd-mquat-benchmark/results/scenarios](https://github.com/Ariyanic/TTC18/tree/master/jastadd-mquat-benchmark/results/scenarios)
The results show that with the same objective, ACO has better execution time than ILP and EMFeR, and with the same execution time, ILP achieves better quality. Also in terms of scalability, by setting time out thredshold on 15 minutes, ILP and ACO solvers could find a valid solution for small-complex-sw and medium benchmarks but EMFeR could not.