We introduce the multitask optimization platform, named MToP, for evolutionary multitasking:
- 40+ multitask evolutionary algorithms for multitask optimization
- 40+ single-task evolutionary algorithms that can handle multitask optimization problems
- 150+ multitask optimization problem cases with real-world applications
- 150+ classical single-task optimization benchmark problems
- 20+ performance metrics covering single- and multi-objective optimization
MToP is a user-friendly tool with a graphical user interface that makes it easy to analyze results, export data, and plot schematics. More importantly, MToP is extensible, allowing users to develop new algorithms and define new problems.
Documents: [Paper - Click Here] / [User Guide - Click Here]
Copyright (c) Yanchi Li. You are free to use the MToP for research purposes. All publications which use this platform should acknowledge the use of "MToP" or "MTO-Platform" and cite as "Y. Li, W. Gong, F. Ming, T. Zhang, S. Li, and Q. Gu, MToP: A MATLAB Optimization Platform for Evolutionary Multitasking, 2023, arXiv:2312.08134"
@Article{Li2023MToP,
title = {{MToP}: A {MATLAB} Optimization Platform for Evolutionary Multitasking},
author = {Yanchi Li and Wenyin Gong and Fei Ming and Tingyu Zhang and Shuijia Li and Qiong Gu},
journal = {arXiv preprint arXiv:2312.08134},
year = {2023},
eprint = {2312.08134},
}
- Fix the bug when GUI parallel runs experiments with save Dec.
- New Algorithm: MTEA-HKTS (Single-objective Multi/Many-task INS24)
- New Problem: Multi-objective sensor coverage problem
- New features:
- Draw dynamic Dec and Obj of populations during optimization in the Test Module
- Pause and Stop buttons can now respond in time by clicking on both the Test and Experiment Module
- Figures sample numbers in the Test Module can be modified, and figures can be exported
- Algorithm and Problem objects can be input in the command line running e.g. "mto(MFEA(), CMT1());"
- New Algorithms:
- CEDA (Constrained Single-objective Multitask SWEC24)
- MTEA-D-TSD (Multi-objective Multitask GECCO24)
- Global-GA (Single-objective Single-task TEVC24)
- KLDE and KLPSO (Single-objective Single-task TEVC23)
- Other classical algorithms: RVEA (MO-ST), SMS-EMOA (MO-ST), IPOP-CMA-ES (SO-ST)
- New Problems:
- Classical Single-Objective Functions with any dimension setting
- Fix some bugs.
- Newly added algorithms:
- MTDE-MKTA (multi-objective multitask TEVC 2024) with application problems
- KR-MTEA (multi/single-objective multitask INS 2023)
- Fix some bugs.
- Newly added algorithms:
- TRADE (single-objective many-task TCYB 2023)
- ASCMFDE (single-objective multitask TEVC 2021)
- Add error value type of WCCI20-MTSO
- Update Operator GA (SBX and polynomial mutation) with more advanced calculation methods. GA-based algorithms now have improved performance.
- The speed of experimental execution is significantly increased, brought by the simultaneous evaluation of whole population decision variables
- 3D task figures of 2-dimensional variables for un-/constrained single-objective multi-/many-/single-task optimization can be plotted in the test module
- Performance metrics can be displayed automatically based on the data type in the experiment module
- Newly added algorithms:
- MKTDE (single-objective multi-task TEVC 2022)
- CCEF-ECHT (constrained single-objective TSMC 2023)
Email: [email protected]
QQ Group: 862974231