Large-scale Sparse Multi-Objective Optimization Research

Python module (mostly Cropover algorithms)

• pymod/main.py: the main class where the simulations are started
• pymod/Cropover.py: the class for the crossover function
• pymod/CropoverTest.py: the class that holds the objective function and the problem generator
• pymod/plotRuns.r: script for plotting results
• pymod/SparseSampler.py: Population initialization sampler.

Matlab module (mostly sparse population algorithms)

• matmod/calcOptimalHVs.m: Calculates the optimal hyper volumes for different reference points for SMOP[1-8] eawef
• matmod/CalHV.m: not my code. Forked from PlatEMO, but modified slight for this project
• matmod/GLOBAL_SPS.m: Extened version of the Global class in PlatEMO that includes information on the final population of the optimization
• matmod/main.m: Main optimization routine for a certain run type (effective or comparative), independent variable (# decision variable or sparsity) and test problem (SMOP[1-8])
• matmod/main_plotting.m: for a given run, plots the different performance metrics over the independent varible of the runs (# number of decision variables or sparsities)
• matmod/main_tables.m: compiles related runs into a single table. Runs are either comparative + decision variables, comparative + sparsity, effective + sparsity, effective + decision variables. Its result will be referred to as a results table
• matmod/main_print_dists.m: Takes a results table and prints distribution of the metrics for those runs
• matmod/main_repetition_summary.m
• matmod/main_significance_tests.m: Takes a results table and runs significant
• testing between SparseEA and NSGA-II/SPS, or with vs without SPS
• matmod/main_testProb_summary.m: Makes statistical summary table of the runs
• matmod/nop.m: Needed for the global object for some reason.
• matmod/quickDemo_2020_10_01.m: quick script comparing NSGA-II/SPS and SparseEA
• matmod/runOpt.m: Wrapper for optimization routine.
• matmod/sparseSampler.m: Sparse sampler source

Data

The data is available from two locations:

• data/Gilgamesh/kroppian/spsRuns on NAS
• /mnt/nas/kroppian/spsRuns on Gilgamesh