Benchmark study of R packages for optimisation

This repository provides R scripts and data to reproduce a comprehensive benchmark study of optimisation algorithms. We use the standard CEC2013 benchmark suite, available as the CRAN package cec2013 0.1, to rigorously evaluate these algorithms. Each benchmark function is tested 10 times in three different dimensions: 30, 50 and 100. The experiment evaluates the following optimisation algorithms: - Genetic Algorithm (geneticAlgorithm using the genalg package). - Differential Evolution (differentialEvolution using the DEoptim package). - Particle Swarm Optimisation (particleSwarm using the pso package). - Convex Partition (convexPartition using the CPoptim package). - Bayesian Treed Constant (tgpCART using the tgp package). - Gaussian Bayesian Optimisation (regr.randomForest using the mlrMBO package). - DIRECT Algorithm (nloptrDirect using the nloptr package). - Nelder-Mead Algorithm (Nelder-Mead using R's built-in optim function).