PYTHON code and MATLAB code for "Multi-fidelity Meta-optimization for Nature Inspired Optimization Algorithms"

Published: 24-07-2020| Version 2 | DOI: 10.17632/pj6d526kzm.2
Hui Li,
Zhiguo Huang


(version 2) We add the MATLAB version ( , hoping researchers who program with MATLAB will find it helpful. The structure of the MATLAB code is: 1. Algorithm (Algorithm.m): 1.1 Basic Algorithm: 1.1.1 PSO.m 1.1.2 GWO.m 2.2 Multi-fidelity Parameter Tuning: 2.2.1 FidelityControlFunction.m 2.2.2 MFOptimizedNIO.m MFOptimizedPSO.m 2.2.3 MFMetaGWO.m 2. Cost Function: 2.1 SphereFunc.m 2.2 CEC14Func.m 2.2.1 input_data 2.2.2 cec14_func.cpp 2.2.3 cec14_func.mexw64 3. Demo: 3.1 DemoMF.m One can run demo as follows: 1. Go into project root: `<YOUR_WORKSPACE>/multi-fidelity-parameter-tuning-matlab` 2. Run the following command in MATLAB window: ``` DemoMF ``` One can compile CEC 2014 as follows: Run the following command to create CEC 2014 library in MATLAB: ``` mex cec14_func.cpp -DWINDOWS ``` ----------- (version 1) The python code is used in the manuscript "Multi-fidelity Meta-optimization for Nature Inspired Optimization Algorithms" submitted to "Applied soft computing". The programming environment is: Python 3.6 or higher. The folders in the package include: 1. algorithms: Basic algorithms, including base class 'Algorithm' and [CS, DE, FOA, GWO, KH, PSO, SSA, WWO, WOA]. 2. applications: An engineering application: source term estimation. 3. benchmarks: Test functions, including base class 'Benchmark', basic test functions and 'CEC2014 Benchmark Suite'. 4. demo: Examples. 5. parameter_tuning: Multi-fidelity meta-NIOs and optimized-NIOs. If you prefer using the command line to run the program, please do not forget to manually add the working directory to 'sys.path'.