| Title: | Genetic Algorithms |
|---|---|
| Description: | Flexible general-purpose toolbox implementing genetic algorithms (GAs) for stochastic optimisation. Binary, real-valued, and permutation representations are available to optimize a fitness function, i.e. a function provided by users depending on their objective function. Several genetic operators are available and can be combined to explore the best settings for the current task. Furthermore, users can define new genetic operators and easily evaluate their performances. Local search using general-purpose optimisation algorithms can be applied stochastically to exploit interesting regions. GAs can be run sequentially or in parallel, using an explicit master-slave parallelisation or a coarse-grain islands approach. For more details see Scrucca (2013) <doi:10.18637/jss.v053.i04> and Scrucca (2017) <doi:10.32614/RJ-2017-008>. |
| Authors: | Luca Scrucca [aut, cre] |
| Maintainer: | Luca Scrucca <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 3.2.5 |
| Built: | 2024-11-11 06:01:04 UTC |
| Source: | https://github.com/luca-scr/ga |
Flexible general-purpose toolbox implementing genetic algorithms (GAs) for stochastic optimisation. Binary, real-valued, and permutation representations are available to optimize a fitness function, i.e. a function provided by users depending on their objective function. Several genetic operators are available and can be combined to explore the best settings for the current task. Furthermore, users can define new genetic operators and easily evaluate their performances. Local search using general-purpose optimisation algorithms can be applied stochastically to exploit interesting regions. GAs can be run sequentially or in parallel, using an explicit master-slave parallelisation or a coarse-grain islands approach.
For a quick intro to GA package see the vignette accompanying the package. Further details are provided in the papers referenced below.
Scrucca L. (2013). GA: A Package for Genetic Algorithms in R. Journal of Statistical Software, 53(4), 1-37, doi:10.18637/jss.v053.i04.
Scrucca, L. (2017) On some extensions to GA package: hybrid optimisation, parallelisation and islands evolution. The R Journal, 9/1, 187-206. doi:10.32614/RJ-2017-008
Luca Scrucca [email protected]
Functions for computing binary to decimal conversion of numbers and vice versa.
decimal2binary(x, length) binary2decimal(x)decimal2binary(x, length) binary2decimal(x)
x |
input value. |
length |
an optional value giving the length of binary string to return. |
decimal2binary converts a numerical value (which is forced to be an integer) to a binary representation, i.e. a vector of 0s and 1s. For real numerical values see the example below.
binary2decimal converts a binary value, i.e. a vector of 0s and 1s, to a decimal representation.
Luca Scrucca
# for integer values dval <- 12 (bval <- decimal2binary(dval)) binary2decimal(bval) # for real values dval <- 12.456 # use (bval <- decimal2binary(dval*1000)) binary2decimal(bval)/1000# for integer values dval <- 12 (bval <- decimal2binary(dval)) binary2decimal(bval) # for real values dval <- 12.456 # use (bval <- decimal2binary(dval*1000)) binary2decimal(bval)/1000
Functions for computing Gray encoding from/to binary strings.
binary2gray(x) gray2binary(x)binary2gray(x) gray2binary(x)
x |
the string to be evaluated |
Gray encoding allows to obtain binary strings not affected by the well-known Hamming cliff problem. With Gray encoding the number of bit differences between any two consecutive values is one, whereas in binary strings this is not always true.
Luca Scrucca
# Consider a five-bit encoding of values 15 and 16 using the standard # binary coding decimal2binary(15, 5) decimal2binary(16, 5) # Moving from 15 to 16 (or vice versa) all five bits need to be changed, # but using Gray encoding the two binary strings differ by one bit. binary2gray(decimal2binary(15, 5)) binary2gray(decimal2binary(16, 5))# Consider a five-bit encoding of values 15 and 16 using the standard # binary coding decimal2binary(15, 5) decimal2binary(16, 5) # Moving from 15 to 16 (or vice versa) all five bits need to be changed, # but using Gray encoding the two binary strings differ by one bit. binary2gray(decimal2binary(15, 5)) binary2gray(decimal2binary(16, 5))
Maximization of a fitness function using Differential Evolution (DE). DE is a population-based evolutionary algorithm for optimisation of fitness functions defined over a continuous parameter space.
de(fitness, lower, upper, popSize = 10*d, stepsize = 0.8, pcrossover = 0.5, ...)de(fitness, lower, upper, popSize = 10*d, stepsize = 0.8, pcrossover = 0.5, ...)
fitness |
the fitness function, any allowable R function which takes as input a vector of values representing a potential solution, and returns a numerical value describing its “fitness”. |
lower |
a vector of length equal to the decision variables providing the lower bounds of the search space. |
upper |
a vector of length equal to the decision variables providing the upper bounds of the search space. |
popSize |
the population size. By default is set at 10 times the number of decision variables. |
pcrossover |
the probability of crossover, by default set to 0.5. |
stepsize |
the stepsize or weighting factor. A value in the interval [0,2], by default set to 0.8. If set at |
... |
additional arguments to be passed to the |
Differential Evolution (DE) is a stochastic evolutionary algorithm that optimises multidimensional real-valued fitness functions without requiring the optimisation problem to be differentiable.
This implimentation follows the description in Simon (2013; Sec. 12.4, and Fig. 12.12) and uses the functionalities available in the ga function for Genetic Algorithms.
The DE selection operator is defined by gareal_de with parameters p = pcrossover and F = stepsize.
Returns an object of class de-class. See de-class for a description of available slots information.
Luca Scrucca [email protected]
Scrucca L. (2013). GA: A Package for Genetic Algorithms in R. Journal of Statistical Software, 53(4), 1-37, doi:10.18637/jss.v053.i04.
Scrucca, L. (2017) On some extensions to GA package: hybrid optimisation, parallelisation and islands evolution. The R Journal, 9/1, 187-206, doi:10.32614/RJ-2017-008.
Simon D. (2013) Evolutionary Optimization Algorithms. John Wiley & Sons.
Price K., Storn R.M., Lampinen J.A. (2005) Differential Evolution: A Practical Approach to Global Optimization. Springer.
summary,de-method,
plot,de-method,
de-class
# 1) one-dimensional function f <- function(x) abs(x)+cos(x) curve(f, -20, 20) DE <- de(fitness = function(x) -f(x), lower = -20, upper = 20) plot(DE) summary(DE) curve(f, -20, 20, n = 1000) abline(v = DE@solution, lty = 3) # 2) "Wild" function, global minimum at about -15.81515 wild <- function(x) 10*sin(0.3*x)*sin(1.3*x^2) + 0.00001*x^4 + 0.2*x + 80 plot(wild, -50, 50, n = 1000) # from help("optim") SANN <- optim(50, fn = wild, method = "SANN", control = list(maxit = 20000, temp = 20, parscale = 20)) unlist(SANN[1:2]) DE <- de(fitness = function(...) -wild(...), lower = -50, upper = 50) plot(DE) summary(DE) # 3) two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) DE <- de(fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50) plot(DE) summary(DE) filled.contour(x1, x2, f, color.palette = bl2gr.colors, plot.axes = { axis(1); axis(2); points(DE@solution, col = "yellow", pch = 3, lwd = 2) }) # 4) two-dimensional Ackley function Ackley <- function(x1, x2) { -20*exp(-0.2*sqrt(0.5*(x1^2 + x2^2))) - exp(0.5*(cos(2*pi*x1) + cos(2*pi*x2))) + exp(1) + 20 } x1 <- x2 <- seq(-3, 3, by = 0.1) f <- outer(x1, x2, Ackley) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) DE <- de(fitness = function(x) -Ackley(x[1], x[2]), lower = c(-3, -3), upper = c(3, 3), stepsize = NA) plot(DE) summary(DE) filled.contour(x1, x2, f, color.palette = bl2gr.colors, plot.axes = { axis(1); axis(2); points(DE@solution, col = "yellow", pch = 3, lwd = 2) }) # 5) Curve fitting example (see Scrucca JSS 2013) ## Not run: # subset of data from data(trees, package = "spuRs") tree <- data.frame(Age = c(2.44, 12.44, 22.44, 32.44, 42.44, 52.44, 62.44, 72.44, 82.44, 92.44, 102.44, 112.44), Vol = c(2.2, 20, 93, 262, 476, 705, 967, 1203, 1409, 1659, 1898, 2106)) richards <- function(x, theta) { theta[1]*(1 - exp(-theta[2]*x))^theta[3] } fitnessL2 <- function(theta, x, y) { -sum((y - richards(x, theta))^2) } DE <- de(fitness = fitnessL2, x = tree$Age, y = tree$Vol, lower = c(3000, 0, 2), upper = c(4000, 1, 4), popSize = 500, maxiter = 1000, run = 100, names = c("a", "b", "c")) summary(DE) ## End(Not run)# 1) one-dimensional function f <- function(x) abs(x)+cos(x) curve(f, -20, 20) DE <- de(fitness = function(x) -f(x), lower = -20, upper = 20) plot(DE) summary(DE) curve(f, -20, 20, n = 1000) abline(v = DE@solution, lty = 3) # 2) "Wild" function, global minimum at about -15.81515 wild <- function(x) 10*sin(0.3*x)*sin(1.3*x^2) + 0.00001*x^4 + 0.2*x + 80 plot(wild, -50, 50, n = 1000) # from help("optim") SANN <- optim(50, fn = wild, method = "SANN", control = list(maxit = 20000, temp = 20, parscale = 20)) unlist(SANN[1:2]) DE <- de(fitness = function(...) -wild(...), lower = -50, upper = 50) plot(DE) summary(DE) # 3) two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) DE <- de(fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50) plot(DE) summary(DE) filled.contour(x1, x2, f, color.palette = bl2gr.colors, plot.axes = { axis(1); axis(2); points(DE@solution, col = "yellow", pch = 3, lwd = 2) }) # 4) two-dimensional Ackley function Ackley <- function(x1, x2) { -20*exp(-0.2*sqrt(0.5*(x1^2 + x2^2))) - exp(0.5*(cos(2*pi*x1) + cos(2*pi*x2))) + exp(1) + 20 } x1 <- x2 <- seq(-3, 3, by = 0.1) f <- outer(x1, x2, Ackley) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) DE <- de(fitness = function(x) -Ackley(x[1], x[2]), lower = c(-3, -3), upper = c(3, 3), stepsize = NA) plot(DE) summary(DE) filled.contour(x1, x2, f, color.palette = bl2gr.colors, plot.axes = { axis(1); axis(2); points(DE@solution, col = "yellow", pch = 3, lwd = 2) }) # 5) Curve fitting example (see Scrucca JSS 2013) ## Not run: # subset of data from data(trees, package = "spuRs") tree <- data.frame(Age = c(2.44, 12.44, 22.44, 32.44, 42.44, 52.44, 62.44, 72.44, 82.44, 92.44, 102.44, 112.44), Vol = c(2.2, 20, 93, 262, 476, 705, 967, 1203, 1409, 1659, 1898, 2106)) richards <- function(x, theta) { theta[1]*(1 - exp(-theta[2]*x))^theta[3] } fitnessL2 <- function(theta, x, y) { -sum((y - richards(x, theta))^2) } DE <- de(fitness = fitnessL2, x = tree$Age, y = tree$Vol, lower = c(3000, 0, 2), upper = c(4000, 1, 4), popSize = 500, maxiter = 1000, run = 100, names = c("a", "b", "c")) summary(DE) ## End(Not run)
An S4 class for differential evolution algorithm
Objects can be created by calls to the de function.
callan object of class "call" representing the matched call;
typea character string specifying the type of genetic algorithm used;
lowera vector providing for each decision variable the lower bounds of the search space in case of real-valued or permutation encoded optimisations. Formerly this slot was named min;
uppera vector providing for each decision variable the upper bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this slot was named max;
namesa vector of character strings providing the names of decision variables (optional);
popSizethe population size;
iterthe actual (or final) iteration of DE search;
runthe number of consecutive generations without any improvement in the best fitness value before the DE is stopped;
maxiterthe maximum number of iterations to run before the DE search is halted;
suggestionsa matrix of user provided solutions and included in the initial population;
populationthe current (or final) population;
elitismthe number of best fitness individuals to survive at each generation;
stepsizethe stepsize or weighting factor;
pcrossoverthe crossover probability;
pmutationthe mutation probability;
optima logical specifying whether or not a local search using general-purpose optimisation algorithms should be used;
fitnessthe values of fitness function for the current (or final) population;
summarya matrix of summary statistics for fitness values at each iteration (along the rows);
bestSolif keepBest = TRUE, the best solutions at each iteration;
fitnessValuethe best fitness value at the final iteration;
solutionthe value(s) of the decision variables giving the best fitness at the final iteration.
Luca Scrucca
For examples of usage see de.
Maximization of a fitness function using genetic algorithms (GAs). Local search using general-purpose optimisation algorithms can be applied stochastically to exploit interesting regions. The algorithm can be run sequentially or in parallel using an explicit master-slave parallelisation.
ga(type = c("binary", "real-valued", "permutation"), fitness, ..., lower, upper, nBits, population = gaControl(type)$population, selection = gaControl(type)$selection, crossover = gaControl(type)$crossover, mutation = gaControl(type)$mutation, popSize = 50, pcrossover = 0.8, pmutation = 0.1, elitism = base::max(1, round(popSize*0.05)), updatePop = FALSE, postFitness = NULL, maxiter = 100, run = maxiter, maxFitness = Inf, names = NULL, suggestions = NULL, optim = FALSE, optimArgs = list(method = "L-BFGS-B", poptim = 0.05, pressel = 0.5, control = list(fnscale = -1, maxit = 100)), keepBest = FALSE, parallel = FALSE, monitor = if(interactive()) gaMonitor else FALSE, seed = NULL)ga(type = c("binary", "real-valued", "permutation"), fitness, ..., lower, upper, nBits, population = gaControl(type)$population, selection = gaControl(type)$selection, crossover = gaControl(type)$crossover, mutation = gaControl(type)$mutation, popSize = 50, pcrossover = 0.8, pmutation = 0.1, elitism = base::max(1, round(popSize*0.05)), updatePop = FALSE, postFitness = NULL, maxiter = 100, run = maxiter, maxFitness = Inf, names = NULL, suggestions = NULL, optim = FALSE, optimArgs = list(method = "L-BFGS-B", poptim = 0.05, pressel = 0.5, control = list(fnscale = -1, maxit = 100)), keepBest = FALSE, parallel = FALSE, monitor = if(interactive()) gaMonitor else FALSE, seed = NULL)
type |
the type of genetic algorithm to be run depending on the nature of decision variables. Possible values are:
|
fitness |
the fitness function, any allowable R function which takes as input an individual |
... |
additional arguments to be passed to the fitness function. This allows to write fitness functions that keep some variables fixed during the search. |
lower |
a vector of length equal to the decision variables providing the lower bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this argument was named |
upper |
a vector of length equal to the decision variables providing the upper bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this argument was named |
nBits |
a value specifying the number of bits to be used in binary encoded optimizations. |
population |
an R function for randomly generating an initial population. See |
selection |
an R function performing selection, i.e. a function which generates a new population of individuals from the current population probabilistically according to individual fitness. See |
crossover |
an R function performing crossover, i.e. a function which forms offsprings by combining part of the genetic information from their parents. See |
mutation |
an R function performing mutation, i.e. a function which randomly alters the values of some genes in a parent chromosome. See |
popSize |
the population size. |
updatePop |
a logical defaulting to |
postFitness |
a user-defined function which, if provided, receives the current |
pcrossover |
the probability of crossover between pairs of chromosomes. Typically this is a large value and by default is set to 0.8. |
pmutation |
the probability of mutation in a parent chromosome. Usually mutation occurs with a small probability, and by default is set to 0.1. |
elitism |
the number of best fitness individuals to survive at each generation. By default the top 5% individuals will survive at each iteration. |
maxiter |
the maximum number of iterations to run before the GA search is halted. |
run |
the number of consecutive generations without any improvement in the best fitness value before the GA is stopped. |
maxFitness |
the upper bound on the fitness function after that the GA search is interrupted. |
names |
a vector of character strings providing the names of decision variables. |
suggestions |
a matrix of solutions strings to be included in the initial population. If provided the number of columns must match the number of decision variables. |
optim |
a logical defaulting to |
optimArgs |
a list controlling the local search algorithm with the following components:
|
keepBest |
a logical argument specifying if best solutions at each iteration should be saved in a slot called |
parallel |
An optional argument which allows to specify if the Genetic Algorithm should be run sequentially or in parallel. For a single machine with multiple cores, possible values are:
In all the cases described above, at the end of the search the cluster is automatically stopped by shutting down the workers. If a cluster of multiple machines is available, evaluation of the fitness function can be executed in parallel using all, or a subset of, the cores available to the machines belonging to the cluster. However, this option requires more work from the user, who needs to set up and register a parallel back end.
In this case the cluster must be explicitly stopped with |
monitor |
a logical or an R function which takes as input the current state of the |
seed |
an integer value containing the random number generator state. This argument can be used to replicate the results of a GA search. Note that if parallel computing is required, the doRNG package must be installed. |
Genetic algorithms (GAs) are stochastic search algorithms inspired by the basic principles of biological evolution and natural selection. GAs simulate the evolution of living organisms, where the fittest individuals dominate over the weaker ones, by mimicking the biological mechanisms of evolution, such as selection, crossover and mutation.
The GA package is a collection of general purpose functions that provide a flexible set of tools for applying a wide range of genetic algorithm methods.
The ga function enables the application of GAs to problems where the decision variables are encoded as "binary", "real-valued", or "permutation" strings.
Default genetic operators are set via gaControl. To retrieve the currently set operators:
gaControl("binary")
gaControl("real-valued")
gaControl("permutation")
Returns an object of class ga-class. See ga-class for a description of available slots information.
Luca Scrucca [email protected]
Back T., Fogel D., Michalewicz Z. (2000). Evolutionary Computation 1: Basic Algorithms and Operators. IOP Publishing Ltd., Bristol and Philadelphia.
Back T., Fogel D., Michalewicz Z. (2000b). Evolutionary Computation 2: Advanced Algorithms and Operators. IOP Publishing Ltd., Bristol and Philadelphia.
Coley D. (1999). An Introduction to Genetic Algorithms for Scientists and Engineers. World Scientific Pub. Co. Inc., Singapore.
Eiben A., Smith J. (2003). Introduction to Evolutionary Computing. Springer-Verlag, Berlin Heidelberg.
Goldberg D. (1989). Genetic Algorithms in Search, Optimization, and Machine Learning. Addison-Wesley Professional, Boston, MA.
Haupt R. L., Haupt S. E. (2004). Practical Genetic Algorithms. 2nd edition. John Wiley & Sons, New York.
Luke S. (2013) Essentials of Metaheuristics, 2nd edition. Lulu.
Scrucca L. (2013). GA: A Package for Genetic Algorithms in R. Journal of Statistical Software, 53(4), 1-37, doi:10.18637/jss.v053.i04.
Scrucca, L. (2017) On some extensions to GA package: hybrid optimisation, parallelisation and islands evolution. The R Journal, 9/1, 187-206, doi:10.32614/RJ-2017-008.
Simon D. (2013) Evolutionary Optimization Algorithms. John Wiley & Sons.
Sivanandam S., Deepa S. (2007). Introduction to Genetic Algorithms. Springer-Verlag, Berlin Heidelberg.
Yu X., Gen M. (2010). Introduction to Evolutionary Algorithms. Springer-Verlag, Berlin Heidelberg.
summary,ga-method,
plot,ga-method,
ga-class,
ga_Population,
ga_Selection,
ga_Crossover,
ga_Mutation,
gaControl.
# 1) one-dimensional function f <- function(x) abs(x)+cos(x) curve(f, -20, 20) fitness <- function(x) -f(x) GA <- ga(type = "real-valued", fitness = fitness, lower = -20, upper = 20) summary(GA) plot(GA) curve(f, -20, 20) abline(v = GA@solution, lty = 3) # 2) one-dimensional function f <- function(x) (x^2+x)*cos(x) # -10 < x < 10 curve(f, -10, 10) # write your own tracing function monitor <- function(obj) { curve(f, -10, 10, main = paste("iteration =", obj@iter)) points(obj@population, obj@fitness, pch = 20, col = 2) rug(obj@population, col = 2) Sys.sleep(0.2) } ## Not run: GA <- ga(type = "real-valued", fitness = f, lower = -10, upper = 10, monitor = monitor) ## End(Not run) # or if you want to suppress the tracing GA <- ga(type = "real-valued", fitness = f, lower = -10, upper = 10, monitor = NULL) summary(GA) monitor(GA) abline(v = GA@solution, lty = 3) # 3) two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) filled.contour(x1, x2, f, color.palette = bl2gr.colors) GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100) summary(GA) plot(GA) # 4) Parallel GA # Simple example of an expensive fitness function obtained artificially by # introducing a pause statement. ## Not run: Rastrigin <- function(x1, x2) { Sys.sleep(0.1) 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } system.time(GA1 <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, monitor = FALSE, seed = 12345)) system.time(GA2 <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, monitor = FALSE, seed = 12345, parallel = TRUE)) ## End(Not run) # 5) Hybrid GA # Example of GA with local search Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, optim = TRUE) summary(GA)# 1) one-dimensional function f <- function(x) abs(x)+cos(x) curve(f, -20, 20) fitness <- function(x) -f(x) GA <- ga(type = "real-valued", fitness = fitness, lower = -20, upper = 20) summary(GA) plot(GA) curve(f, -20, 20) abline(v = GA@solution, lty = 3) # 2) one-dimensional function f <- function(x) (x^2+x)*cos(x) # -10 < x < 10 curve(f, -10, 10) # write your own tracing function monitor <- function(obj) { curve(f, -10, 10, main = paste("iteration =", obj@iter)) points(obj@population, obj@fitness, pch = 20, col = 2) rug(obj@population, col = 2) Sys.sleep(0.2) } ## Not run: GA <- ga(type = "real-valued", fitness = f, lower = -10, upper = 10, monitor = monitor) ## End(Not run) # or if you want to suppress the tracing GA <- ga(type = "real-valued", fitness = f, lower = -10, upper = 10, monitor = NULL) summary(GA) monitor(GA) abline(v = GA@solution, lty = 3) # 3) two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20, col.palette = bl2gr.colors) filled.contour(x1, x2, f, color.palette = bl2gr.colors) GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100) summary(GA) plot(GA) # 4) Parallel GA # Simple example of an expensive fitness function obtained artificially by # introducing a pause statement. ## Not run: Rastrigin <- function(x1, x2) { Sys.sleep(0.1) 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } system.time(GA1 <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, monitor = FALSE, seed = 12345)) system.time(GA2 <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, monitor = FALSE, seed = 12345, parallel = TRUE)) ## End(Not run) # 5) Hybrid GA # Example of GA with local search Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 100, optim = TRUE) summary(GA)
Functions implementing crossover genetic operator.
ga_spCrossover(object, parents, ...) gabin_spCrossover(object, parents, ...) gabin_uCrossover(object, parents, ...) gareal_spCrossover(object, parents, ...) gareal_waCrossover(object, parents, ...) gareal_laCrossover(object, parents, ...) gareal_blxCrossover(object, parents, a = 0.5, ...) gareal_laplaceCrossover(object, parents, a = 0, b = 0.15, ...) gaperm_cxCrossover(object, parents, ...) gaperm_pmxCrossover(object, parents, ...) gaperm_oxCrossover(object, parents, ...) gaperm_pbxCrossover(object, parents, ...)ga_spCrossover(object, parents, ...) gabin_spCrossover(object, parents, ...) gabin_uCrossover(object, parents, ...) gareal_spCrossover(object, parents, ...) gareal_waCrossover(object, parents, ...) gareal_laCrossover(object, parents, ...) gareal_blxCrossover(object, parents, a = 0.5, ...) gareal_laplaceCrossover(object, parents, a = 0, b = 0.15, ...) gaperm_cxCrossover(object, parents, ...) gaperm_pmxCrossover(object, parents, ...) gaperm_oxCrossover(object, parents, ...) gaperm_pbxCrossover(object, parents, ...)
object |
An object of class |
parents |
A two-rows matrix of values indexing the parents from the current population. |
... |
Further arguments passed to or from other methods. |
a, b
|
Parameters of genetic operators. |
Return a list with two elements:
children |
a matrix of dimension 2 times the number of decision variables containing the generated offsprings; |
fitness |
a vector of length 2 containing the fitness values for the offsprings. A value |
Luca Scrucca
Functions implementing mutation genetic operator.
gabin_raMutation(object, parent, ...) gareal_raMutation(object, parent, ...) gareal_nraMutation(object, parent, ...) gareal_rsMutation(object, parent, ...) gareal_powMutation(object, parent, pow = 10, ...) gaperm_simMutation(object, parent, ...) gaperm_ismMutation(object, parent, ...) gaperm_swMutation(object, parent, ...) gaperm_dmMutation(object, parent, ...) gaperm_scrMutation(object, parent, ...)gabin_raMutation(object, parent, ...) gareal_raMutation(object, parent, ...) gareal_nraMutation(object, parent, ...) gareal_rsMutation(object, parent, ...) gareal_powMutation(object, parent, pow = 10, ...) gaperm_simMutation(object, parent, ...) gaperm_ismMutation(object, parent, ...) gaperm_swMutation(object, parent, ...) gaperm_dmMutation(object, parent, ...) gaperm_scrMutation(object, parent, ...)
object |
An object of class |
parent |
A vector of values for the parent from the current population where mutation should occur. |
... |
Further arguments passed to or from other methods. |
pow |
Parameters of genetic operators. |
Return a vector of values containing the mutated string.
Luca Scrucca
A function which calculates the mutation probability for the current iteration. This enables to use GAs with variable mutation rate (see examples).
ga_pmutation(object, p0 = 0.5, p = 0.01, T = round(object@maxiter/2), ...)ga_pmutation(object, p0 = 0.5, p = 0.01, T = round(object@maxiter/2), ...)
object |
An object of class |
p0 |
initial probability of mutation. |
p |
limiting probability of mutation. |
T |
maximum iteration after which it should converges to |
... |
Further arguments passed to or from other methods. |
Return a numeric value in the range (0,1).
Luca Scrucca
## Not run: Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 500, run = 100, pmutation = ga_pmutation) plot(GA) GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 500, run = 100, pmutation = function(...) ga_pmutation(..., p0 = 0.1)) plot(GA) ## End(Not run)## Not run: Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 500, run = 100, pmutation = ga_pmutation) plot(GA) GA <- ga(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 50, maxiter = 500, run = 100, pmutation = function(...) ga_pmutation(..., p0 = 0.1)) plot(GA) ## End(Not run)
Functions for creating a random initial population to be used in genetic algorithms.
gabin_Population(object, ...) gareal_Population(object, ...) gaperm_Population(object, ...)gabin_Population(object, ...) gareal_Population(object, ...) gaperm_Population(object, ...)
object |
An object of class |
... |
Further arguments passed to or from other methods. |
gabin_Population generates a random population of object@nBits binary values;
gareal_Population generates a random (uniform) population of real values in the range [object@min, object@max];
gaperm_Population generates a random (uniform) population of integer values in the range [object@min, object@max].
Return a matrix of dimension object@popSize times the number of decision variables.
Luca Scrucca
Functions implementing selection genetic operator.
ga_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) ga_nlrSelection(object, q = 0.25, ...) ga_rwSelection(object, ...) ga_tourSelection(object, k = 3, ...) gabin_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gabin_nlrSelection(object, q = 0.25, ...) gabin_rwSelection(object, ...) gabin_tourSelection(object, k = 3, ...) gareal_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gareal_nlrSelection(object, q = 0.25, ...) gareal_rwSelection(object, ...) gareal_tourSelection(object, k = 3, ...) gareal_lsSelection(object, ...) gareal_sigmaSelection(object, ...) gaperm_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gaperm_nlrSelection(object, q = 0.25, ...) gaperm_rwSelection(object, ...) gaperm_tourSelection(object, k = 3, ...) gareal_de(object, F = 0.8, p = 0.5, ...)ga_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) ga_nlrSelection(object, q = 0.25, ...) ga_rwSelection(object, ...) ga_tourSelection(object, k = 3, ...) gabin_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gabin_nlrSelection(object, q = 0.25, ...) gabin_rwSelection(object, ...) gabin_tourSelection(object, k = 3, ...) gareal_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gareal_nlrSelection(object, q = 0.25, ...) gareal_rwSelection(object, ...) gareal_tourSelection(object, k = 3, ...) gareal_lsSelection(object, ...) gareal_sigmaSelection(object, ...) gaperm_lrSelection(object, r = 2/(object@popSize * (object@popSize - 1)), q = 2/object@popSize, ...) gaperm_nlrSelection(object, q = 0.25, ...) gaperm_rwSelection(object, ...) gaperm_tourSelection(object, k = 3, ...) gareal_de(object, F = 0.8, p = 0.5, ...)
object |
An object of class |
r |
A tuning parameter for the GA selection operator. |
q |
A tuning parameter for the GA selection operator. |
k |
A tuning parameter for the GA selection operator. |
F, p
|
Tuning parameters for the DE selection operator. |
... |
Further arguments passed to or from other methods. |
Return a list with two elements:
population |
a matrix of dimension |
fitness |
a vector of length |
Luca Scrucca
An S4 class for genetic algorithms
Objects can be created by calls to the ga function.
callan object of class "call" representing the matched call;
typea character string specifying the type of genetic algorithm used;
lowera vector providing for each decision variable the lower bounds of the search space in case of real-valued or permutation encoded optimisations. Formerly this slot was named min;
uppera vector providing for each decision variable the upper bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this slot was named max;
nBitsa value specifying the number of bits to be used in binary encoded optimizations;
namesa vector of character strings providing the names of decision variables (optional);
popSizethe population size;
iterthe actual (or final) iteration of GA search;
runthe number of consecutive generations without any improvement in the best fitness value before the GA is stopped;
maxiterthe maximum number of iterations to run before the GA search is halted;
suggestionsa matrix of user provided solutions and included in the initial population;
populationthe current (or final) population;
elitismthe number of best fitness individuals to survive at each generation;
pcrossoverthe crossover probability;
pmutationthe mutation probability;
optima logical specifying whether or not a local search using general-purpose optimisation algorithms should be used;
fitnessthe values of fitness function for the current (or final) population;
summarya matrix of summary statistics for fitness values at each iteration (along the rows);
bestSolif keepBest = TRUE, the best solutions at each iteration;
fitnessValuethe best fitness value at the final iteration;
solutionthe value(s) of the decision variables giving the best fitness at the final iteration.
Luca Scrucca
For examples of usage see ga.
Default settings for genetic operators used in the GA package.
gaControl(...)gaControl(...)
... |
no arguments, a single character vector, or a named list with components. |
If the function is called with no arguments returns the current default settings, i.e., a list with the following default components:
binarypopulation = "gabin_Population"
selection = "gabin_lrSelection"
crossover = "gabin_spCrossover"
mutation = "gabin_raMutation"
real-valuedpopulation = "gareal_Population"
selection = "gareal_lsSelection"
crossover = "gareal_laCrossover"
mutation = "gareal_raMutation"
permutationpopulation = "gaperm_Population"
selection = "gaperm_lrSelection"
crossover = "gaperm_oxCrossover"
mutation = "gaperm_simMutation"
epsthe tolerance value used by the package functions. By default set at sqrt(.Machine$double.eps).
useRcppa logical specifying if a faster C++ implementation of genetic operators should be used (TRUE by default), or an R implementation.
The function may be called with a single string specifying the name of the component. In this case the function returns the current default settings.
To change the default values, a named component must be followed by a single value (in case of eps or useRcpp) or a list of component(s) specifying the name of the function for a genetic operator. See the Examples section.
If the argument list is empty the function returns the current list of values. If the argument list is not empty, the returned list is invisible.
The parameter values set via a call to this function will remain in effect for the rest of the session, affecting the subsequent behaviour of the functions for which the given parameters are relevant.
Luca Scrucca
# get and save defaults defaultControl <- gaControl() print(defaultControl) # get current defaults only for binary search gaControl("binary") # set defaults for selection operator of binary search gaControl("binary" = list(selection = "gabin_tourSelection")) gaControl("binary") # set defaults for selection and crossover operators of binary search gaControl("binary" = list(selection = "ga_rwSelection", crossover = "gabin_uCrossover")) gaControl("binary") # restore defaults gaControl(defaultControl) gaControl()# get and save defaults defaultControl <- gaControl() print(defaultControl) # get current defaults only for binary search gaControl("binary") # set defaults for selection operator of binary search gaControl("binary" = list(selection = "gabin_tourSelection")) gaControl("binary") # set defaults for selection and crossover operators of binary search gaControl("binary" = list(selection = "ga_rwSelection", crossover = "gabin_uCrossover")) gaControl("binary") # restore defaults gaControl(defaultControl) gaControl()
Maximization of a fitness function using islands genetic algorithms (ISLGAs). This is a distributed multiple-population GA, where the population is partitioned into several subpopulations and assigned to separated islands. Independent GAs are executed in each island, and only occasionally sparse exchanges of individuals are performed among the islands. In principle islands can evolve sequentially, but increased computational efficiency is obtained by running GAs in parallel on each island. The latter is called island parallel GAs (ISLPGAs) and it is used by default.
gaisl(type = c("binary", "real-valued", "permutation"), fitness, ..., lower, upper, nBits, population = gaControl(type)$population, selection = gaControl(type)$selection, crossover = gaControl(type)$crossover, mutation = gaControl(type)$mutation, popSize = 100, numIslands = 4, migrationRate = 0.1, migrationInterval = 10, pcrossover = 0.8, pmutation = 0.1, elitism = base::max(1, round(popSize/numIslands*0.05)), updatePop = FALSE, postFitness = NULL, maxiter = 1000, run = maxiter, maxFitness = Inf, names = NULL, suggestions = NULL, optim = FALSE, optimArgs = list(method = "L-BFGS-B", poptim = 0.05, pressel = 0.5, control = list(fnscale = -1, maxit = 100)), parallel = TRUE, monitor = if(interactive()) gaislMonitor else FALSE, seed = NULL)gaisl(type = c("binary", "real-valued", "permutation"), fitness, ..., lower, upper, nBits, population = gaControl(type)$population, selection = gaControl(type)$selection, crossover = gaControl(type)$crossover, mutation = gaControl(type)$mutation, popSize = 100, numIslands = 4, migrationRate = 0.1, migrationInterval = 10, pcrossover = 0.8, pmutation = 0.1, elitism = base::max(1, round(popSize/numIslands*0.05)), updatePop = FALSE, postFitness = NULL, maxiter = 1000, run = maxiter, maxFitness = Inf, names = NULL, suggestions = NULL, optim = FALSE, optimArgs = list(method = "L-BFGS-B", poptim = 0.05, pressel = 0.5, control = list(fnscale = -1, maxit = 100)), parallel = TRUE, monitor = if(interactive()) gaislMonitor else FALSE, seed = NULL)
type |
the type of genetic algorithm to be run depending on the nature of decision variables. Possible values are:
|
fitness |
the fitness function, any allowable R function which takes as input an individual |
... |
additional arguments to be passed to the fitness function. This allows to write fitness functions that keep some variables fixed during the search. |
lower |
a vector of length equal to the decision variables providing the lower bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this argument was named |
upper |
a vector of length equal to the decision variables providing the upper bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this argument was named |
nBits |
a value specifying the number of bits to be used in binary encoded optimizations. |
population |
an R function for randomly generating an initial population. See |
numIslands |
an integer value specifying the number of islands to be used in a ring topology, in which each island is connected unidirectionally with another island, hence forming a single continuous pathway. |
migrationRate |
a value in the range $[0,1]$ providing the proportion of individuals that should migrate between the islands. |
migrationInterval |
an integer value specifying the number of iterations at which exchange of individuals takes place. |
selection |
an R function performing selection, i.e. a function which generates a new population of individuals from the current population probabilistically according to individual fitness. See |
crossover |
an R function performing crossover, i.e. a function which forms offsprings by combining part of the genetic information from their parents. See |
mutation |
an R function performing mutation, i.e. a function which randomly alters the values of some genes in a parent chromosome. See |
popSize |
the population size. |
updatePop |
a logical defaulting to |
postFitness |
a user-defined function which, if provided, receives the current |
pcrossover |
the probability of crossover between pairs of chromosomes. Typically this is a large value and by default is set to 0.8. |
pmutation |
the probability of mutation in a parent chromosome. Usually mutation occurs with a small probability, and by default is set to 0.1. |
elitism |
the number of best fitness individuals to survive at each generation. By default the top 5% individuals in each island will survive at each iteration. |
maxiter |
the maximum number of iterations to run before the GA search is halted. |
run |
the number of consecutive generations without any improvement in the best fitness value before the GA is stopped. |
maxFitness |
the upper bound on the fitness function after that the GA search is interrupted. |
names |
a vector of character strings providing the names of decision variables. |
suggestions |
a matrix of solutions strings to be included in the initial population. If provided the number of columns must match the number of decision variables. |
optim |
a logical defaulting to |
optimArgs |
a list controlling the local search algorithm with the following components:
|
parallel |
An optional argument which allows to specify if the Islands Genetic Algorithm should be run sequentially or in parallel. For a single machine with multiple cores, possible values are:
In all the cases described above, at the end of the search the cluster is automatically stopped by shutting down the workers. If a cluster of multiple machines is available, evolution of GAs on each island can be executed in parallel using all, or a subset of, the cores available to the machines belonging to the cluster. However, this option requires more work from the user, who needs to set up and register a parallel back end.
In this case the cluster must be explicitly stopped with |
monitor |
a logical or an R function which takes as input the current state of the |
seed |
an integer value containing the random number generator state. This argument can be used to replicate the results of a ISLGA search. Note that if parallel computing is required, the doRNG package must be installed. |
Genetic algorithms (GAs) are stochastic search algorithms inspired by the basic principles of biological evolution and natural selection. GAs simulate the evolution of living organisms, where the fittest individuals dominate over the weaker ones, by mimicking the biological mechanisms of evolution, such as selection, crossover and mutation.
The gaisl function implements the islands GAs approach, where the population is partitioned into several subpopulations and assigned to separated islands. Independent GAs are executed in each island, and only occasionally sparse exchanges of individuals are performed among the islands. The algorithm can be run in parallel or sequentially.
For more information on GAs see ga.
Returns an object of class gaisl-class. See gaisl-class for a description of available slots information.
Luca Scrucca [email protected]
Luque G., Alba E. (2011) Parallel Genetic Algorithms: Theory and Real World Applications. Springer.
Luke S. (2013) Essentials of Metaheuristics, 2nd edition. Lulu.
Scrucca, L. (2017) On some extensions to GA package: hybrid optimisation, parallelisation and islands evolution. The R Journal, 9/1, 187-206. https://journal.r-project.org/archive/2017/RJ-2017-008/
summary,gaisl-method,
plot,gaisl-method,
gaisl-class,
ga
## Not run: # two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20) filled.contour(x1, x2, f, color.palette = jet.colors) GA <- gaisl(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 80, maxiter = 500, numIslands = 4, migrationInterval = 50) summary(GA) plot(GA) ## End(Not run)## Not run: # two-dimensional Rastrigin function Rastrigin <- function(x1, x2) { 20 + x1^2 + x2^2 - 10*(cos(2*pi*x1) + cos(2*pi*x2)) } x1 <- x2 <- seq(-5.12, 5.12, by = 0.1) f <- outer(x1, x2, Rastrigin) persp3D(x1, x2, f, theta = 50, phi = 20) filled.contour(x1, x2, f, color.palette = jet.colors) GA <- gaisl(type = "real-valued", fitness = function(x) -Rastrigin(x[1], x[2]), lower = c(-5.12, -5.12), upper = c(5.12, 5.12), popSize = 80, maxiter = 500, numIslands = 4, migrationInterval = 50) summary(GA) plot(GA) ## End(Not run)
An S4 class for islands genetic algorithms (ISLGAs)
Objects can be created by calls to the gaisl function.
callan object of class "call" representing the matched call;
typea character string specifying the type of genetic algorithm used;
lowera vector providing for each decision variable the lower bounds of the search space in case of real-valued or permutation encoded optimisations. Formerly this slot was named min;
uppera vector providing for each decision variable the upper bounds of the search space in case of real-valued or permutation encoded optimizations. Formerly this slot was named max;
nBitsa value specifying the number of bits to be used in binary encoded optimizations;
namesa vector of character strings providing the names of decision variables (optional);
popSizethe population size;
numIslandsthe number of islands;
migrationRatethe migration rate;
migrationIntervalthe migration interval;
maxiterthe maximum number of ISLGA iterations before the search is halted;
runthe number of consecutive generations without any improvement in the best fitness value before the ISLGA is stopped;
maxiterthe maximum number of iterations to run before the GA search is halted;
suggestionsa matrix of user provided solutions and included in the initial population;
elitismthe number of best fitness individuals to survive at each generation;
pcrossoverthe crossover probability;
pmutationthe mutation probability;
optima logical specifying whether or not a local search using general-purpose optimisation algorithms should be used;
islandsa list containing the objects of class ga corresponding to each island GA evolution;
summarya list of matrices of summary statistics for fitness values at each iteration (along the rows). Each element of the list corresponds to the evolution of an island;
fitnessValuesa list of best fitness values found in each island at the final iteration;
solutionsa list of matrices, one for each island, containing the values of the decision variables giving the best fitness at the final iteration;
fitnessValuethe best fitness value at the final iteration;
solutiona matrix containing the values of the decision variables giving the best fitness at the final iteration.
Luca Scrucca
For examples of usage see gaisl.
Functions to print summary statistics of fitness values at each iteration of a GA search.
gaMonitor(object, digits = getOption("digits"), ...) gaislMonitor(object, digits = getOption("digits"), ...)gaMonitor(object, digits = getOption("digits"), ...) gaislMonitor(object, digits = getOption("digits"), ...)
object |
an object of class |
digits |
minimal number of significant digits. |
... |
further arguments passed to or from other methods. |
These functions print a summary of current GA step on the console.
By default, gaMonitor is called in interactive sessions by ga. The old monitoring function, used as the default until version 2.2 of GA package, is provided in gaMonitor2.
By default, gaislMonitor is called in interactive sessions by gaisl.
Luca Scrucca
A function which returns fitness summary statistics at each iteration of GA search.
gaSummary(x, ...)gaSummary(x, ...)
x |
a vector of fitness values for which summary statistics should be computed. |
... |
further arguments passed to or from other methods. |
This function computes summary statistics for a vector of fitness values at current iteration of GA search.
A vector with the following values: (max, mean, median, min)
Luca Scrucca
The class "numericOrNA" is a simple class union
(setClassUnion) of "numeric" and "logical".
Since it is a virtual Class, no objects may be created from it.
showClass("numericOrNA")showClass("numericOrNA")
Functions for creating a vector of colours from pre-specified palettes.
jet.colors(n) spectral.colors(n) bl2gr.colors(n)jet.colors(n) spectral.colors(n) bl2gr.colors(n)
n |
a numerical value specifying the number of colours in the palette. |
jet.colors() creates a palette of colours which tend to have high brightness and not uniform luminance. Furthermore, the brightest colours, yellow and cyan, are used for intermediate data values, and this has the effect of emphasizing uninteresting (and arbitrary) values while de-emphasizing the extremes. For these reasons this popular palette is not recommended.spectral.colors() creates a palette based on ColorBrewer https://colorbrewer2.org, so the resulting colours have a much uniform luminance.
The bl2gr.colors() palette returns a palette of colours from blue to green.
Returns a character vector of colours encoded in hexadecimal values.
jet.colors(9) spectral.colors(9) bl2gr.colors(9) par(mfrow = c(3,1), mar = c(1,1,1,1)) n = 21 image(1:21, 1, as.matrix(1:21), col = jet.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n") image(1:21, 1, as.matrix(1:21), col = spectral.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n") image(1:21, 1, as.matrix(1:21), col = bl2gr.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n")jet.colors(9) spectral.colors(9) bl2gr.colors(9) par(mfrow = c(3,1), mar = c(1,1,1,1)) n = 21 image(1:21, 1, as.matrix(1:21), col = jet.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n") image(1:21, 1, as.matrix(1:21), col = spectral.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n") image(1:21, 1, as.matrix(1:21), col = bl2gr.colors(21), ylab = "", xlab = "", xaxt = "n", yaxt = "n", bty = "n")
ga-class.A method for obtaining the names of parameters or decision variables
from an object of class ga-class.
parNames(object, ...) ## S4 method for signature 'ga' parNames(object, ...)parNames(object, ...) ## S4 method for signature 'ga' parNames(object, ...)
object |
An object of class |
... |
Further arguments, currently not used. |
A list of character values providing the names of parameters or decision variables.
Luca Scrucca
This function draws a perspective plot of a surface with different levels in different colours.
persp3D(x, y, z, theta = 30, phi = 20, d = 5, expand = 2/3, xlim = range(x, finite = TRUE), ylim = range(y, finite = TRUE), zlim = range(z, finite = TRUE), levels = pretty(zlim, nlevels), nlevels = 20, col.palette = jet.colors, border = NA, ticktype = "detailed", xlab = NULL, ylab = NULL, zlab = NULL, ...)persp3D(x, y, z, theta = 30, phi = 20, d = 5, expand = 2/3, xlim = range(x, finite = TRUE), ylim = range(y, finite = TRUE), zlim = range(z, finite = TRUE), levels = pretty(zlim, nlevels), nlevels = 20, col.palette = jet.colors, border = NA, ticktype = "detailed", xlab = NULL, ylab = NULL, zlab = NULL, ...)
x, y
|
locations of grid lines at which the values in |
z |
a matrix containing the values to be plotted (NAs are allowed). |
theta, phi
|
angles defining the viewing direction. |
d |
a value which can be used to vary the strength of the perspective transformation. |
expand |
a expansion factor applied to the |
xlim, ylim, zlim
|
x-, y- and z-limits for the axes. |
levels |
a vector of values specifying the levels to be used for plotting the surface with different colours. |
nlevels |
a value specifying the number of levels to be used for plotting. This value is used if |
col.palette |
the colour palette used for plotting. |
border |
the colour of the line drawn around the surface facets. By default is set to |
ticktype |
a character specifying the type of axes tickmarks. By default |
xlab, ylab, zlab
|
character strings specifying the titles for the axes. |
... |
Further arguments passed to the function |
This function enhances the default perspective plot for drawing 3-dimensional surfaces.
Return a list with the following elements:
persp |
the viewing transformation matrix (see |
levels |
a vector of values giving the levels used for plotting the surface; |
colors |
a vector of strings giving the colour used for plotting the surface. |
Luca Scrucca
y <- x <- seq(-10, 10, length=60) f <- function(x,y) { r <- sqrt(x^2+y^2); 10 * sin(r)/r } z <- outer(x, y, f) persp3D(x, y, z, theta = 30, phi = 30, expand = 0.5) persp3D(x, y, z, col.palette = heat.colors, phi = 30, theta = 225, box = TRUE, border = NA, shade = .4) x1 <- seq(-3,3,length=50) x2 <- seq(-3,3,length=50) y <- function(x1, x2) sin(x1)+cos(x2) persp3D(x1, x2, outer(x1,x2,y), zlab="y", theta = 150, phi = 20, expand = 0.6)y <- x <- seq(-10, 10, length=60) f <- function(x,y) { r <- sqrt(x^2+y^2); 10 * sin(r)/r } z <- outer(x, y, f) persp3D(x, y, z, theta = 30, phi = 30, expand = 0.5) persp3D(x, y, z, col.palette = heat.colors, phi = 30, theta = 225, box = TRUE, border = NA, shade = .4) x1 <- seq(-3,3,length=50) x2 <- seq(-3,3,length=50) y <- function(x1, x2) sin(x1)+cos(x2) persp3D(x1, x2, outer(x1,x2,y), zlab="y", theta = 150, phi = 20, expand = 0.6)
The plot method for de-class objects gives a plot
of best and average fitness values found during the iterations of the
DE search.
## S4 method for signature 'de' plot(x, y, ylim, cex.points = 0.7, col = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1)), pch = c(16, 1), lty = c(1,2), legend = TRUE, grid = graphics::grid, ...)## S4 method for signature 'de' plot(x, y, ylim, cex.points = 0.7, col = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1)), pch = c(16, 1), lty = c(1,2), legend = TRUE, grid = graphics::grid, ...)
x |
An object of class |
y |
Not used. |
ylim |
A vector of two values specifying the limits on the y-axis. |
cex.points |
The magnification to be used for points. |
col |
The colours to be used for best and average fitness values. |
pch |
The type of points to be used for best and average fitness values. |
lty |
The type of lines to be used for best and average fitness values. |
legend |
A logical specifying if a legend should be included. |
grid |
A function for grid drawing of NULL to avoid drawing one. |
... |
Further arguments, currently not used. |
Plot best and average fitness values at each iteration of DE search.
The method invisibly return a data.frame with the iterations and summary statistics for the fitness function evaluated at each iteration.
Luca Scrucca
# See examples in help(de)# See examples in help(de)
The plot method for ga-class objects gives a plot
of best and average fitness values found during the iterations of the
GA search.
## S4 method for signature 'ga' plot(x, y, ylim, cex.points = 0.7, col = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1)), pch = c(16, 1), lty = c(1,2), legend = TRUE, grid = graphics::grid, ...)## S4 method for signature 'ga' plot(x, y, ylim, cex.points = 0.7, col = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1)), pch = c(16, 1), lty = c(1,2), legend = TRUE, grid = graphics::grid, ...)
x |
An object of class |
y |
Not used. |
ylim |
A vector of two values specifying the limits on the y-axis. |
cex.points |
The magnification to be used for points. |
col |
The colours to be used for best and average fitness values. |
pch |
The type of points to be used for best and average fitness values. |
lty |
The type of lines to be used for best and average fitness values. |
legend |
A logical specifying if a legend should be included. |
grid |
A function for grid drawing of NULL to avoid drawing one. |
... |
Further arguments, currently not used. |
Plot best and average fitness values at each iteration of GA search.
The method invisibly return a data.frame with the iterations and summary statistics for the fitness function evaluated at each iteration.
Luca Scrucca
# See examples in help(ga) # The following code shows how to obtain graphs using the # ggplot2 plotting system ## Not run: GA <- ga(type = "real-valued", fitness = function(x) -(abs(x)+cos(x)), lower = -20, upper = 20, popSize = 20, pmutation = 0.2, maxiter = 50) out <- plot(GA) library(reshape2) df <- melt(out[,c(1:3,5)], id.var = "iter") library(ggplot2) ggplot(out) + geom_ribbon(aes(x = iter, ymin = median, ymax = max, colour = "median", fill = "median")) + geom_line(aes(x = iter, y = max, colour = "max")) + geom_point(aes(x = iter, y = max, colour = "max")) + geom_line(aes(x = iter, y = mean, colour = "mean"), lty = 2) + geom_point(aes(x = iter, y = mean, colour = "mean"), pch = 1) + xlab("Generation") + ylab("Fitness values") + scale_colour_manual(breaks = c("max", "mean", "median"), values = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1))) + scale_fill_manual(breaks = "median", values = adjustcolor("green3", alpha.f = 0.1)) + guides(fill = "none", colour = guide_legend(override.aes = list(fill = c(NA, NA, adjustcolor("green3", alpha.f = 0.1)), pch = c(19,1,NA)))) + theme_bw() + theme(legend.title = element_blank(), legend.pos = "top", legend.background = element_blank()) ## End(Not run)# See examples in help(ga) # The following code shows how to obtain graphs using the # ggplot2 plotting system ## Not run: GA <- ga(type = "real-valued", fitness = function(x) -(abs(x)+cos(x)), lower = -20, upper = 20, popSize = 20, pmutation = 0.2, maxiter = 50) out <- plot(GA) library(reshape2) df <- melt(out[,c(1:3,5)], id.var = "iter") library(ggplot2) ggplot(out) + geom_ribbon(aes(x = iter, ymin = median, ymax = max, colour = "median", fill = "median")) + geom_line(aes(x = iter, y = max, colour = "max")) + geom_point(aes(x = iter, y = max, colour = "max")) + geom_line(aes(x = iter, y = mean, colour = "mean"), lty = 2) + geom_point(aes(x = iter, y = mean, colour = "mean"), pch = 1) + xlab("Generation") + ylab("Fitness values") + scale_colour_manual(breaks = c("max", "mean", "median"), values = c("green3", "dodgerblue3", adjustcolor("green3", alpha.f = 0.1))) + scale_fill_manual(breaks = "median", values = adjustcolor("green3", alpha.f = 0.1)) + guides(fill = "none", colour = guide_legend(override.aes = list(fill = c(NA, NA, adjustcolor("green3", alpha.f = 0.1)), pch = c(19,1,NA)))) + theme_bw() + theme(legend.title = element_blank(), legend.pos = "top", legend.background = element_blank()) ## End(Not run)
The plot method for gaisl-class objects gives a plot
of best fitness values found in each island during the GA iterations.
## S4 method for signature 'gaisl' plot(x, y, ...)## S4 method for signature 'gaisl' plot(x, y, ...)
x |
An object of class |
y |
Not used. |
... |
Further arguments passed to |
Plot best fitness values found in each island during the GA iterations.
The method invisibly return a list with the following components:
iter |
a vector of values specifying the iteration. |
summary |
a matrix of best fitness values for each island along the columns. |
Luca Scrucca
# See examples in help(gaisl)# See examples in help(gaisl)
Summary method for class de-class.
## S4 method for signature 'de' summary(object, ...) ## S3 method for class 'summary.de' print(x, digits = getOption("digits"), ...)## S4 method for signature 'de' summary(object, ...) ## S3 method for class 'summary.de' print(x, digits = getOption("digits"), ...)
object |
an object of class |
x |
an object of class |
digits |
number of significant digits. |
... |
further arguments passed to or from other methods. |
The summary function returns an object of class summary.de which can be printed by the corresponding print method. The function also returns invisibly a list with the information from the differential evolution search.
Luca Scrucca
f <- function(x) abs(x)+cos(x) DE <- de(fitness = function(x) -f(x), lower = -20, upper = 20, run = 50) out <- summary(DE) print(out) str(out)f <- function(x) abs(x)+cos(x) DE <- de(fitness = function(x) -f(x), lower = -20, upper = 20, run = 50) out <- summary(DE) print(out) str(out)
Summary method for class ga-class.
## S4 method for signature 'ga' summary(object, ...) ## S3 method for class 'summary.ga' print(x, digits = getOption("digits"), ...)## S4 method for signature 'ga' summary(object, ...) ## S3 method for class 'summary.ga' print(x, digits = getOption("digits"), ...)
object |
an object of class |
x |
an object of class |
digits |
number of significant digits. |
... |
further arguments passed to or from other methods. |
The summary function returns an object of class summary.ga which can be printed by the corresponding print method. The function also returns invisibly a list with the information from the genetic algorithm search.
Luca Scrucca
f <- function(x) abs(x)+cos(x) GA <- ga(type = "real-valued", fitness = function(x) -f(x), lower = -20, upper = 20, run = 50) out <- summary(GA) print(out) str(out)f <- function(x) abs(x)+cos(x) GA <- ga(type = "real-valued", fitness = function(x) -f(x), lower = -20, upper = 20, run = 50) out <- summary(GA) print(out) str(out)
Summary method for class gaisl-class.
## S4 method for signature 'gaisl' summary(object, ...) ## S3 method for class 'summary.gaisl' print(x, digits = getOption("digits"), ...)## S4 method for signature 'gaisl' summary(object, ...) ## S3 method for class 'summary.gaisl' print(x, digits = getOption("digits"), ...)
object |
an object of class |
x |
an object of class |
digits |
number of significant digits. |
... |
further arguments passed to or from other methods. |
The summary function returns an object of class summary.gaisl which can be printed by the corresponding print method. The function also returns invisibly a list with the information from the islands genetic algorithm search.
Luca Scrucca
## Not run: f <- function(x) abs(x)+cos(x) GA <- gaisl(type = "real-valued", fitness = function(x) -f(x), lower = -20, upper = 20, run = 10, numIslands = 4) out <- summary(GA) print(out) str(out) ## End(Not run)## Not run: f <- function(x) abs(x)+cos(x) GA <- gaisl(type = "real-valued", fitness = function(x) -f(x), lower = -20, upper = 20, run = 10, numIslands = 4) out <- summary(GA) print(out) str(out) ## End(Not run)