| Type: | Package |
| Title: | Evolutionary Parameter Estimation for 'Repast Simphony' Models |
| Version: | 0.5.0 |
| Date: | 2018-08-30 |
| Author: | Antonio Prestes Garcia [aut, cre], Alfonso Rodriguez-Paton [aut, ths] |
| Maintainer: | Antonio Prestes Garcia <antonio.pgarcia@alumnos.upm.es> |
| URL: | https://github.com/antonio-pgarcia/evoper |
| BugReports: | https://github.com/antonio-pgarcia/evoper/issues |
| Description: | The EvoPER, Evolutionary Parameter Estimation for Individual-based Models is an extensible package providing optimization driven parameter estimation methods using metaheuristics and evolutionary computation techniques (Particle Swarm Optimization, Simulated Annealing, Ant Colony Optimization for continuous domains, Tabu Search, Evolutionary Strategies, ...) which could be more efficient and require, in some cases, fewer model evaluations than alternatives relying on experimental design. Currently there are built in support for models developed with 'Repast Simphony' Agent-Based framework (https://repast.github.io/) and with NetLogo (https://ccl.northwestern.edu/netlogo/) which are the most used frameworks for Agent-based modeling. |
| License: | MIT + file LICENSE |
| LazyData: | TRUE |
| Depends: | rrepast |
| Imports: | methods, futile.logger, boot, reshape, ggplot2, deSolve, plot3D, plyr, data.table, utils, RNetLogo |
| RoxygenNote: | 6.0.1 |
| Suggests: | testthat |
| NeedsCompilation: | no |
| Packaged: | 2018-08-30 17:50:41 UTC; antonio.pgarcia |
| Repository: | CRAN |
| Date/Publication: | 2018-08-30 23:20:06 UTC |
Estimates
Description
A simple class for encapsulating the return of metaheuristic methods
Magnitude
Description
Calculates the magnitude order for a given value
Usage
Magnitude(v)
Arguments
v |
The numerical value |
Value
The magnitude order
NLWrapper.FindJar
Description
Search for the netlogo jar file on the provided path
Usage
NLWrapper.FindJar(path)
Arguments
path |
The base path for searching |
Value
The path for NetLogo jar file
NLWrapper.GetParameter
Description
Gets the value of a model parameter
Usage
NLWrapper.GetParameter(obj, name)
Arguments
obj |
The object retuned by NLWrapper.Model |
name |
The parameter name string or the collection of parameter names |
Value
The parameter values
Examples
## Not run:
rm(list=ls())
p<- "C:/Program Files/NetLogo 6.0.4/app"
m<- file.path(nlpath, "models", "Sample Models", "Biology", "Wolf Sheep Predation.nlogo")
o<- NLWrapper.Model(p, m)
v<- NLWrapper.GetParameter(o, c("initial-number-sheep"))
or
v<- NLWrapper.GetParameter(o, c("initial-number-sheep","initial-number-wolves")))
## End(Not run)
NLWrapper.Model
Description
This wrapper prepares the environment and instantiates the model
Usage
NLWrapper.Model(netlogodir, modelfile, dataset, maxtime)
Arguments
netlogodir |
The base path of NetLogo installation |
modelfile |
The absolute path for NetLogo model file |
dataset |
The names of model variables |
maxtime |
The total number of iterations |
Examples
## Not run:
rm(list=ls())
p<- "C:/Program Files/NetLogo 6.0.4/app"
output<- c("count sheep", "count wolves")
m<- file.path(nlpath, "models", "Sample Models", "Biology", model, "Wolf Sheep Predation.nlogo")
o<- NLWrapper.Model(p, m, output, 150)
## End(Not run)
NLWrapper.Run
Description
Executes a NetLogo Model using rNetLogo
Usage
NLWrapper.Run(obj, r = 1, seed = c())
Arguments
obj |
The object retuned by NLWrapper.Model |
r |
The number of replications |
seed |
The collection of random seeds |
NLWrapper.RunExperiment
Description
Executes a NetLogo Model using rNetLogo
Usage
NLWrapper.RunExperiment(obj, r = 1, design, FUN)
Arguments
obj |
The object retuned by NLWrapper.Model |
r |
The number of replications |
design |
The desing matrix holding parameter sampling |
FUN |
THe calibration function. |
Value
A list containing the the parameters, the calibration functio output and the whole resultset
Examples
## Not run:
rm(list=ls())
objectivefn<- function(params, results) { 0 }
f<- AddFactor(name="initial-number-sheep",min=100,max=250)
f<- AddFactor(factors=f, name="initial-number-wolves",min=50,max=150)
f<- AddFactor(factors=f, name="grass-regrowth-time",min=30,max=100)
f<- AddFactor(factors=f, name="sheep-gain-from-food",min=1,max=50)
f<- AddFactor(factors=f, name="wolf-gain-from-food",min=1,max=100)
f<- AddFactor(factors=f, name="sheep-reproduce",min=1,max=20)
f<- AddFactor(factors=f, name="wolf-reproduce",min=1,max=20)
design<- AoE.LatinHypercube(factors=f)
p<- "C:/Program Files/NetLogo 6.0.4/app"
m<- file.path(p, "models", "Sample Models", "Biology", "Wolf Sheep Predation.nlogo")
output<- c("count sheep", "count wolves")
o<- NLWrapper.Model(p, m, output, 150)
v<- RunExperiment(o, r=1, design, objectivefn)
NLWrapper.Shutdown(o)
## End(Not run)
NLWrapper.SetParameter
Description
Set parameter values
Usage
NLWrapper.SetParameter(obj, parameters)
Arguments
obj |
The object retuned by NLWrapper.Model |
parameters |
The data frame containing the paramters |
Examples
## Not run:
rm(list=ls())
p<- "C:/Program Files/NetLogo 6.0.4/app"
m<- file.path(nlpath, "models", "Sample Models", "Biology", "Wolf Sheep Predation.nlogo")
o<- NLWrapper.Model(p, m)
## End(Not run)
NLWrapper.SetRandomSeed
Description
Configures the random seed
Usage
NLWrapper.SetRandomSeed(obj, seed)
Arguments
obj |
The object retuned by NLWrapper.Model |
seed |
The new random seed |
NLWrapper.Shutdown
Description
This wrapper terminates RNetLogo execution environment
Usage
NLWrapper.Shutdown(obj)
Arguments
obj |
The object retuned by NLWrapper.Model |
NetLogoFunction
Description
NetLogoFunction class
ObjectiveFunction class
Description
The base class for optimization functions.
Fields
objectThe raw output of objective function
objectiveThe objective function
parametersThe parameter list for objective function
valueThe results from objective function
Options
Description
The base class for the options for the optimization metaheuristics
Fields
typeThe configuration type
neighborhoodThe neighborhood function for population methods
discreteFlag indicating that and specific algorithm is discrete or continuous
nlevelzDefault value for generating parameter levels when range is provided, default value is 5
containerThe object holding the configuration otions
OptionsACOR
Description
Options for ACOR method
OptionsEES1
Description
Options for EvoPER Evolutionary Stratety 1
OptionsEES2
Description
Options for Serial Dilutions method
Fields
dilutionsThe desired dilutions
OptionsFactory
Description
Instantiate the Options class required for the specific metaheuristic method.
Usage
OptionsFactory(type, v = NULL)
Arguments
type |
The metaheuristic method |
v |
The options object |
Value
Options object
OptionsPSO
Description
Options for PSO optimization metaheuristic
OptionsSAA
Description
Options for SAA method
Fields
temperatureThe temperature dacay function
OptionsTS
Description
Options for Tabu search optimization metaheuristic
PlainFunction
Description
PlainFunction Class
RepastFunction
Description
RepastFunction class
Ant colony optimization for continuous domains
Description
An implementation of Ant Colony Optimization algorithm for continuous variables.
Usage
abm.acor(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("acor", f)
## End(Not run)
EvoPER Evolutionary Strategy 1
Description
This function tries to provide a rough approximation to best solution when no information is available for the correct range of input parameters for the objective function. It can useful for studying the behavior of individual-based models with high variability in the output variables showing nonlinear behaviors.
Usage
abm.ees1(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("ees1", f)
## End(Not run)
EvoPER Evolutionary Strategy 2
Description
This function tries to provide a rough approximation to best solution when no information is available for the correct range of input parameters for the objective function. It can useful for studying the behavior of individual-based models with high variability in the output variables showing nonlinear behaviors.
Usage
abm.ees2(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("ees2", f)
## End(Not run)
abm.pso
Description
An implementaion of Particle Swarm Optimization method for parameter estimation of Individual-based models.
Usage
abm.pso(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
References
[1] Kennedy, J., & Eberhart, R. (1995). Particle swarm optimization. In Proceedings of ICNN 95 - International Conference on Neural Networks (Vol. 4, pp. 1942-1948). IEEE.
[2] Poli, R., Kennedy, J., & Blackwell, T. (2007). Particle swarm optimization. Swarm Intelligence, 1(1), 33-57.
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("pso", f)
## End(Not run)
abm.saa
Description
An implementation of Simulated Annealing Algorithm optimization method for parameter estimation of Individual-based models.
Usage
abm.saa(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
Value
The best solution.
References
[1] Kirkpatrick, S., Gelatt, C. D., & Vecchi, M. P. (1983). Optimization by Simulated Annealing. Science, 220(4598).
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("saa", f)
## End(Not run)
## Not run:
## A Repast defined function
f<- RepastFunction$new("/usr/models/BactoSim(HaldaneEngine-1.0)","ds::Output",300)
## or a plain function
f1<- function(x1,x2,x3,x4) {
10 * (x1 - 1)^2 + 20 * (x2 - 2)^2 + 30 * (x3 - 3)^2 + 40 * (x4 - 4)^2
}
f<- PlainFunction$new(f1)
f$addFactor(name="cyclePoint",min=0,max=90)
f$addFactor(name="conjugationCost",min=0,max=100)
f$addFactor(name="pilusExpressionCost",min=0,max=100)
f$addFactor(name="gamma0",min=1,max=10)
abm.saa(f, 100, 1, 100, 0.75)
## End(Not run)
Tabu Search metaheuristic
Description
An implementation of Tabu Search algorithm for parameter estimation
Usage
abm.tabu(objective, options = NULL)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
References
[1] Fred Glover (1989). "Tabu Search - Part 1". ORSA Journal on Computing, 190-206. doi:10.1287/ijoc.1.3.190. [2] Fred Glover (1990). "Tabu Search - Part 2". ORSA Journal on Computing, 4-32. doi:10.1287/ijoc.2.1.4.
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
or
f$Parameter0(name="x1",levels=c(0:4))
f$Parameter0(name="x2",levels=c(-2,-1,0,1,2))
extremize("tabu", f)
## End(Not run)
acor.F
Description
Helper function for extracting the 'F' function evaluations from archive ACOr 'T'
Usage
acor.F(T)
Arguments
T |
The solution archive |
Value
The F matrix
acor.N
Description
Helper function for getting the size of solution
Usage
acor.N(T)
Arguments
T |
The solution archive |
Value
The size 'n' of a solution 's'
acor.S
Description
Helper function for extracting solution 'S' from archive 'T'
Usage
acor.S(T)
Arguments
T |
The solution archive |
Value
The solution matrix
acor.W
Description
Helper function for extracting the 'W' function evaluations from archive ACOr 'T'
Usage
acor.W(T)
Arguments
T |
The solution archive |
Value
The weight vector
acor.archive
Description
This function is used for creating and maintaining the ACOr archive 'T'. The function keeps the track of 'k' solotion in the archive.
Usage
acor.archive(s, f, w, k, T = NULL)
Arguments
s |
The solution 'ants' |
f |
The evaluation of solution |
w |
The weight vector |
k |
The archive size |
T |
The current archive |
Value
The solution archive
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
Select the lth gaussian function
Description
Given a weight vector calculate the probabilities of selecting the lth gaussian function and return the index of lht gaussian selected with probability p
Usage
acor.lthgaussian(W)
Arguments
W |
The vector of weights |
Value
The index of lht gaussian function
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
Gaussian kernel choosing probability
Description
Calculate the probability of choosing the lth Gaussian function
Usage
acor.probabilities(W, l = NULL)
Arguments
W |
The vector of weights |
l |
The lth element of algorithm solution archive T |
Value
The vector of probabilities 'p'
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
Sigma calculation for ACOr
Description
Calculate the value of sigma
Usage
acor.sigma(Xi, k, T)
Arguments
Xi |
The algorithm parameter |
k |
The solution archive size |
T |
The solution archive |
Value
The sigma value
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
acor.updateants
Description
Update the solution using the gaussian kernel
Usage
acor.updateants(S, N, W, t.mu, t.sigma)
Arguments
S |
The current solution ants |
N |
The numnber of required ants in solution |
W |
The weight vector |
t.mu |
The 'mean' from solution archive |
t.sigma |
The value of sigma from solution archive |
Value
The new solution ants
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
Weight calculation for ant colony optimization
Description
Calculates the weight element of ACOr algorithm for the solution archive.
Usage
acor.weigth(q, k, l)
Arguments
q |
The Algorithm parameter. When small best-ranked solution is preferred |
k |
The Archive size |
l |
The lth element of algorithm solution archive T |
Value
A scalar or a vector with calculated weigth.
References
[1] Socha, K., & Dorigo, M. (2008). Ant colony optimization for continuous domains. European Journal of Operational Research, 185(3), 1155-1173. http://doi.org/10.1016/j.ejor.2006.06.046
assert
Description
The assert function stop the execution if the logical
expression given by the parameter expresion is false.
Usage
assert(expresion, string)
Arguments
expresion |
Some logical expression |
string |
The text message to show if expression does not hold |
bestFitness
Description
Given a set S of N solutions created with sortSolution, this function returns the fitness component fot the best solution.
Usage
bestFitness(S)
Arguments
S |
The solution set |
Value
The best fitness value
bestSolution
Description
Given a set S of N solutions created with sortSolution, this function returns the best solution found.
Usage
bestSolution(S)
Arguments
S |
The solution set |
Value
The best solution
cbuf
Description
Simple implementation of a circular buffer.
Usage
cbuf(b, v, e)
Arguments
b |
The variable holding the current buffer content |
v |
The new valued to be added to b |
e |
The length of circular buffer |
Value
The buffer b plus the element v minus the least recently added element
compare.algorithms1
Description
Compare the number of function evalutions and convergence for the following optimization algorithms, ("saa","pso","acor","ees1").
Usage
compare.algorithms1(F, seeds = c(27, 2718282, 36190727, 3141593, -91190721,
-140743, 1321))
Arguments
F |
The function to be tested |
seeds |
The random seeds which will be used for testing algorithms |
Examples
## Not run:
rm(list=ls())
d.cigar4<- compare.algorithms1(f0.cigar4)
d.schaffer4<- compare.algorithms1(f0.schaffer4)
d.griewank4<- compare.algorithms1(f0.griewank4)
d.bohachevsky4<- compare.algorithms1(f0.bohachevsky4)
d.rosenbrock4<- compare.algorithms1(f0.rosenbrock4)
## End(Not run)
contourplothelper
Description
Simple helper function for countour plots
Usage
contourplothelper(d, x, y, z, nbins = 32, binwidth = c(10, 10),
points = c(300, 300), title = NULL)
Arguments
d |
A data frame. |
x |
A string with the dataframe column name for x axis. |
y |
A string with the dataframe column name for y axis. |
z |
A string with the dataframe column name for z axis. |
nbins |
The number bins. The default is 32. |
binwidth |
The binwidths for 'kde2d'. Can be an scalar or a vector. |
points |
The number of grid points. Can be an scalar or a vector. |
title |
The optional plot title. May be omited. |
ees1.challenge
Description
Repeat the evalution of best solution to tacke with variability.
Usage
ees1.challenge(solution, objective)
Arguments
solution |
The Problem solution |
objective |
The objective function |
ees1.explore
Description
Explore the solution space on the neighborhood of solution 's' in order to find a new best.
Usage
ees1.explore(s, weight, p = 0.01)
Arguments
s |
The Problem solution |
weight |
The exploration intensity |
p |
The mutation probability |
ees1.mating
Description
This function 'mix' the elements present in the solution. The parameter 'mu' controls the intensity of mixing. Low values give preference to best solution components and high values make the values being select randomly.
Usage
ees1.mating(solution, mu)
Arguments
solution |
The Problem solution |
mu |
The mixing intensity ratio, from 0 to 1. The mix intensity controls de the probability of chosing a worst solutions |
ees1.mating1
Description
This function 'mix' the elements present in the solution. The parameter 'mu' controls the intensity of mixing. Low values give preference to best solution components and high values make the values being select randomly.
Usage
ees1.mating1(solution, mu)
Arguments
solution |
The Problem solution |
mu |
The mixing intensity ratio, from 0 to 1. The mix intensity controls de the probability of chosing a worst solutions |
ees1.mutation
Description
Performs the mutation on generated solution
Usage
ees1.mutation(solution, mates, p = 0.01)
Arguments
solution |
The Problem solution |
mates |
The mixed parents |
p |
The mutation probability |
ees1.recombination
Description
Performs the recombination on solution
Usage
ees1.recombination(solution, mates)
Arguments
solution |
The Problem solution |
mates |
The mixed parents |
ees.selection
Description
Select the elements with best fitness but accept uphill moves with probability 'kkappa'.
Usage
ees1.selection(s0, s1, kkappa)
Arguments
s0 |
The current best solution set |
s1 |
The new solution |
kkappa |
The selection pressure |
elog.debug
Description
Wrapper for logging debug messages.
Usage
elog.debug(...)
Arguments
... |
Variable number of arguments including a format string. |
elog.error
Description
Wrapper for logging error messages.
Usage
elog.error(...)
Arguments
... |
Variable number of arguments including a format string. |
elog.info
Description
Wrapper for logging info messages.
Usage
elog.info(...)
Arguments
... |
Variable number of arguments including a format string. |
elog.level
Description
Configure the current log level
Usage
elog.level(level = NULL)
Arguments
level |
The log level (ERROR|WARN|INFO|DEBUG) |
Value
The log level
enforceBounds
Description
Checks if parameters fall within upper an lower bounds
Usage
enforceBounds(particles, factors)
Arguments
particles |
The particle set |
factors |
the defined range for objective function parameters |
Value
The particle inside the valid limits
es.evaluate
Description
For each element in solution 's' evaluate the respective fitness.
Usage
es.evaluate(f, s, enforce = TRUE)
Arguments
f |
A reference to an instance of objective function |
s |
The set of solutions |
enforce |
If true the values are enforced to fall within provided range |
Value
The solution ordered by its fitness.
extremize
Description
Entry point for optimization functions
Usage
extremize(type, objective, options = NULL)
Arguments
type |
The optimization method (aco,pso,saa,sda) |
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
options |
An apropiate instance from a sublclass of Options class |
Examples
## Not run:
f<- PlainFunction$new(f0.rosenbrock2)
f$Parameter(name="x1",min=-100,max=100)
f$Parameter(name="x2",min=-100,max=100)
extremize("pso", f)
## End(Not run)
f0.ackley
Description
The ackley function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0. Domain xi E [-32.768, 32.768], for all i = 1, ..., d
Usage
f0.ackley(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
https://www.sfu.ca/~ssurjano/ackley.html
f0.ackley4
Description
The ackley function of four variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.ackley4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.adtn.rosenbrock2
Description
Two variable Rosenbrock function with random additive noise.
Usage
f0.adtn.rosenbrock2(x1, x2)
Arguments
x1 |
Parameter 1 |
x2 |
Parameter 2 |
f0.bohachevsky
Description
The Bohachevsky function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.bohachevsky(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.bohachevsky4
Description
The Bohachevsky function of four variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.bohachevsky4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.cigar
Description
The Cigar function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.cigar(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.cigar4
Description
The Cigar function of four variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.cigar4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.griewank
Description
The griewank function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.griewank(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.griewank4
Description
The griewank function of four variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.griewank4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.nlnn.rosenbrock2
Description
Two variable Rosenbrock function with random additive noise.
Usage
f0.nlnn.rosenbrock2(x1, x2)
Arguments
x1 |
Parameter 1 |
x2 |
Parameter 2 |
Period tuning for Predator-Prey base
Description
This function is an example on how EvoPER can be used for estimating the parameter values in order to produce oscilations with the desired period. It is not intended to be used directelly, the provided wrappers should be instead.
Usage
f0.periodtuningpp(x1, x2, x3, x4, period)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
period |
The desired oscilation period |
Value
The solution fitness cost
Period tuning of 12 time units for Predator-Prey
Description
This function is an example on how EvoPER can be used for estimating the parameter values in order to produce oscilations with the desired period.
Usage
f0.periodtuningpp12(x1, x2, x3, x4)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
Value
The solution fitness cost
Examples
## Not run:
rm(list=ls())
set.seed(-27262565)
f<- PlainFunction$new(f0.periodtuningpp12)
f$Parameter(name="x1",min=0.5,max=2)
f$Parameter(name="x2",min=0.5,max=2)
f$Parameter(name="x3",min=0.5,max=2)
f$Parameter(name="x4",min=0.5,max=2)
extremize("pso", f)
## End(Not run)
Period tuning of 24 time units for Predator-Prey
Description
This function is an example on how EvoPER can be used for estimating the parameter values in order to produce oscilations with the desired period.
Usage
f0.periodtuningpp24(x1, x2, x3, x4)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
Value
The solution fitness cost
Examples
## Not run:
rm(list=ls())
set.seed(-27262565)
f<- PlainFunction$new(f0.periodtuningpp24)
f$Parameter(name="x1",min=0.5,max=2)
f$Parameter(name="x2",min=0.5,max=2)
f$Parameter(name="x3",min=0.5,max=2)
f$Parameter(name="x4",min=0.5,max=2)
extremize("pso", f)
## End(Not run)
Period tuning of 48 time units for Predator-Prey
Description
This function is an example on how EvoPER can be used for estimating the parameter values in order to produce oscilations with the desired period.
Usage
f0.periodtuningpp48(x1, x2, x3, x4)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
Value
The solution fitness cost
Examples
## Not run:
rm(list=ls())
set.seed(-27262565)
f<- PlainFunction$new(f0.periodtuningpp24)
f$Parameter(name="x1",min=0.5,max=2)
f$Parameter(name="x2",min=0.5,max=2)
f$Parameter(name="x3",min=0.5,max=2)
f$Parameter(name="x4",min=0.5,max=2)
extremize("pso", f)
## End(Not run)
Period tuning of 72 time units for Predator-Prey
Description
This function is an example on how EvoPER can be used for estimating the parameter values in order to produce oscilations with the desired period.
Usage
f0.periodtuningpp72(x1, x2, x3, x4)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
Value
The solution fitness cost
Examples
## Not run:
rm(list=ls())
set.seed(-27262565)
f<- PlainFunction$new(f0.periodtuningpp24)
f$Parameter(name="x1",min=0.5,max=2)
f$Parameter(name="x2",min=0.5,max=2)
f$Parameter(name="x3",min=0.5,max=2)
f$Parameter(name="x4",min=0.5,max=2)
extremize("pso", f)
## End(Not run)
f0.rosenbrock2
Description
Two variable Rosenbrock function, where f(1,1) = 0
Usage
f0.rosenbrock2(x1, x2)
Arguments
x1 |
Parameter 1 |
x2 |
Parameter 2 |
f0.rosenbrock4
Description
The rosenbrock function of 4 variables for testing optimization methods. The global optima for the function is given by xi = 1, forall i E 1...N, f(x) = 0.
Usage
f0.rosenbrock4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.rosenbrockn
Description
The rosenbrock function of N variables for testing optimization methods. The global optima for the function is given by xi = 1, forall i E 1...N, f(x) = 0.
Usage
f0.rosenbrockn(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.schaffer
Description
The schaffer function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.schaffer(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.schaffer4
Description
The Schaffer function of four variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f0.schaffer4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.schwefel
Description
The schwefel function of N variables for testing optimization methods. The global optima for the function is given by xi = 420.96874636, forall i E 1...N, f(x) = 0. The range of xi is [-500,500]
Usage
f0.schwefel(...)
Arguments
... |
The variadic list of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f0.schwefel4
Description
The schwefel function of N variables for testing optimization methods. The global optima for the function is given by xi = 420.96874636, forall i E 1...N, f(x) = 0. The range of xi is [-500,500]
Usage
f0.schwefel4(x1, x2, x3, x4)
Arguments
x1 |
The first function variable |
x2 |
The second function variable |
x3 |
The third function variable |
x4 |
The fourth function variable |
Value
The function value
f0.test
Description
Simple test function f(1,2,3,4) = 0
Usage
f0.test(x1, x2, x3, x4)
Arguments
x1 |
Parameter 1 |
x2 |
Parameter 2 |
x3 |
Parameter 3 |
x4 |
Parameter 4 |
f1.ackley
Description
The ackley function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0. Domain xi E [-32.768, 32.768], for all i = 1, ..., d
Usage
f1.ackley(x)
Arguments
x |
The vector of function parameters |
Value
The function value
References
https://www.sfu.ca/~ssurjano/ackley.html
f1.adtn.rosenbrock2
Description
Two variable Rosenbrock function with random additive noise.
Usage
f1.adtn.rosenbrock2(x)
Arguments
x |
Parameter vector |
f1.bohachevsky
Description
The Bohachevsky function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f1.bohachevsky(x)
Arguments
x |
The vector of function parameters |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.cigar
Description
The Cigar function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f1.cigar(x)
Arguments
x |
The vector of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.griewank
Description
The griewank function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f1.griewank(x)
Arguments
x |
The vector of function parameters |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.nlnn.rosenbrock2
Description
Two variable Rosenbrock function with random additive noise.
Usage
f1.nlnn.rosenbrock2(x)
Arguments
x |
Parameter vector |
f1.rosenbrock2
Description
Two variable Rosenbrock function, where f(c(1,1)) = 0
Usage
f1.rosenbrock2(x)
Arguments
x |
Parameter vector |
f1.rosenbrockn
Description
The rosenbrock function of N variables for testing optimization methods. The global optima for the function is given by xi = 1, forall i E 1...N, f(x) = 0.
Usage
f1.rosenbrockn(x)
Arguments
x |
The vector of function parameters |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.schaffer
Description
The schaffer function of N variables for testing optimization methods. The global optima for the function is given by xi = 0, forall i E 1...N, f(x) = 0.
Usage
f1.schaffer(x)
Arguments
x |
The vector of function parameters |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.schwefel
Description
The schwefel function of N variables for testing optimization methods. The global optima for the function is given by xi = 420.96874636, forall i E 1...N, f(x) = 0. The range of xi is [-500,500]
Usage
f1.schwefel(x)
Arguments
x |
The vector of function variables. |
Value
The function value
References
http://deap.gel.ulaval.ca/doc/dev/api/benchmarks.html
f1.test
Description
Simple test function f(c(1,2,3,4)) = 0
Usage
f1.test(x)
Arguments
x |
Parameter vector |
fixdfcolumns
Description
Coerce dataframe columns to a specic type.
Usage
fixdfcolumns(df, cols = c(), skip = TRUE, type = as.numeric)
Arguments
df |
The data frame. |
cols |
The dataframe columns to be skiped or included. |
skip |
If TRUE the column names in 'cols' are skiped. When FALSE logic is inverted. |
type |
The type for which data frame columns must be converted. |
Value
The data frame with converted column types.
generateSolution
Description
Generates a problema solution using discrete leves
Usage
generateSolution(parameters, size)
Arguments
parameters |
The Objective Function parameter list |
size |
The solution size |
Value
The solution set
getFitness
Description
Given a set S of N solutions created with sortSolution, this function returns the solution component fot the best solution.
Usage
getFitness(S, i = NULL)
Arguments
S |
The solution set |
i |
The fitness index, if null return the whole column. |
Value
The selected fitness entry
getSolution
Description
Given a set S of N solutions created with sortSolution, this function returns the solution component. A solutions is a set of solutions and their associated fitness
Usage
getSolution(S)
Arguments
S |
The solution set |
Value
The solution set
gm.mean
Description
Simple implementation for geometric mean
Usage
gm.mean(x)
Arguments
x |
data |
Value
geometric mean for data
gm.sd
Description
Simple implementation for geometric standard deviation
Usage
gm.sd(x, mu = NULL)
Arguments
x |
data |
mu |
The geometric mean. If not provided it is calculated. |
Value
geometric standard deviation for data
histplothelper
Description
Simple helper for ploting histograms
Usage
histplothelper(d, x, title = NULL)
Arguments
d |
A data frame. |
x |
A string with the dataframe column name for histogram |
title |
The plot title |
Value
A ggplot2 plot object
initSolution
Description
Creates the initial Solution population taking into account the lower an upper bounds of provided experiment factors.
Usage
initSolution(parameters, N = 20, sampling = "mcs")
Arguments
parameters |
The Objective Function parameter list |
N |
The size of Solution population |
sampling |
The population sampling scheme, namelly <mcs|lhs|ffs> standing respectively for montecarlo sampling, latin hypercube sampling and full factorial sampling |
Value
A random set of solutions
lowerBound
Description
Checks if parameters is greater than the lower bounds
Usage
lowerBound(particles, factors)
Arguments
particles |
The particle set |
factors |
the defined range for objective function parameters |
Value
The particle greater than or equal to lower limit
naiveperiod
Description
A naive approach for finding the period in a series of data points
Usage
naiveperiod(d)
Arguments
d |
The data to search period |
Value
A list with the average period and amplitude
paramconverter
Description
Convert parameter from continuous to discrete and vice-versa if needed
Usage
paramconverter(parameters, discrete, levelz = 5)
Arguments
parameters |
The current parameter set |
discrete |
The desired parameter type |
levelz |
When discrete is true the number of levels to be generated |
Value
The parameter collection casted to desired mode
partSolutionSpace
Description
Creates the initial Solution population taking into account the lower an upper bounds of provided experiment factors. This method works by dividing the solution space into partitions of size 'd' and then creating a full factorial combination of partitions.
Usage
partSolutionSpace(parameters, d = 4)
Arguments
parameters |
The Objective Function parameter list |
d |
The partition size. Default value 4. |
Value
A set of solutions
pop.first
Description
pop an element
Usage
pop.first(x)
Arguments
x |
The element collection |
Value
The first element added to list FIFO
pop.last
Description
pop an element
Usage
pop.last(x)
Arguments
x |
The element collection |
Value
The last element added to list LIFO
predatorprey
Description
The solver for Lotka-Volterra differential equation.
Usage
predatorprey(x1, x2, x3, x4)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effecto on predator |
Value
The ODE solution
predatorprey.plot0
Description
Generate a plot for the predator-prey ODE output.
Usage
predatorprey.plot0(x1, x2, x3, x4, title = NULL)
Arguments
x1 |
The growth rate of prey |
x2 |
The decay rate of predator |
x3 |
The predating effect on prey |
x4 |
The predating effect on predator |
title |
The optional plot title. May be omited. |
Value
An ggplot2 object
Examples
## Not run:
predatorprey.plot0(1.351888, 1.439185, 1.337083, 0.9079049)
## End(Not run)
predatorprey.plot1
Description
Simple wrapper for 'predatorprey.plot0' accepting the parameters as a list.
Usage
predatorprey.plot1(x, title = NULL)
Arguments
x |
A list containing the values of predator/prey parameters c1, c2, c3 and c4 denoting respectivelly the growth rate of prey, the decay rate of predator, the predating effect on prey and the predating effect on predator |
title |
The optional plot title. May be omited. |
Value
An ggplot2 object
Examples
## Not run:
rm(list=ls())
predatorprey.plot1(v$getBest()[1:4])
## End(Not run)
pso.velocity
Description
Calculates the PSO Velocity
Usage
pso.Velocity(W = 1, Vi, phi1, phi2, Pi, Pg, Xi)
Arguments
W |
Weight (Inertia weight or constriction coefficient) |
Vi |
Current Velocity vector |
phi1 |
Acceleration coefficient toward the previous best |
phi2 |
Acceleration coefficient toward the global best |
Pi |
Personal best |
Pg |
Neighborhood best |
Xi |
Particle vector |
Value
Updated velocity
pso.best
Description
Search for the best particle solution which minimize the objective function.
Usage
pso.best(objective, particles)
Arguments
objective |
The results of evaluating the objective function |
particles |
The particles tested |
Value
The best particle
pso.chi
Description
Implementation of constriction coefficient
Usage
pso.chi(phi1, phi2)
Arguments
phi1 |
Acceleration coefficient toward the previous best |
phi2 |
Acceleration coefficient toward the global best |
Value
The calculated constriction coefficient
pso.lbest
Description
Finds the lbest for the particle 'i' using the topology function given by the topology parameter.
Usage
pso.lbest(i, pbest, topology)
Arguments
i |
The particle position |
pbest |
The pbest particle collection |
topology |
The desired topology function |
Value
The lbes for i th particle
pso.neighborhood.K2
Description
The neighborhood function for a simple linear topology where every particle has k = 2 neighbors
Usage
pso.neighborhood.K2(i, n)
Arguments
i |
The particle position |
n |
the size of particle population |
pso.neighborhood.K4
Description
The von neumann neighborhood function for a lattice-based topology where every particle has k = 4 neighbors
Usage
pso.neighborhood.K4(i, n)
Arguments
i |
The particle position |
n |
the size of particle population |
pso.neighborhood.KN
Description
Simple helper method for 'gbest' neighborhood
Usage
pso.neighborhood.KN(i, n)
Arguments
i |
The particle position |
n |
the size of particle population |
pso.printbest
Description
Shows the best particle of each of simulated generations
Usage
pso.printbest(objective, particles, generation, title)
Arguments
objective |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
particles |
The current particle population |
generation |
The current generation |
title |
Some informational text to be shown |
push
Description
push an element
Usage
push(x, v)
Arguments
x |
The collection of elements |
v |
The value to be pushed |
Value
The collection of elements
random.whell
Description
A simple randon seed generator
Usage
random.wheel()
Value
A random number for seeding
saa.bolt
Description
Temperature function boltzmann
Usage
saa.bolt(t0, k)
Arguments
t0 |
The current temperature |
k |
The annealing value |
Value
The new temperature
saa.neighborhood
Description
Generates neighbor solutions for simulated annealing
Usage
saa.neighborhood(f, S, d, n)
Arguments
f |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
S |
The current solution to find a neighbor |
d |
The distance from current solution S |
n |
The number of parameters to be perturbed |
Value
The neighbor of solution S
saa.neighborhood1
Description
Generates neighbor solutions perturbing one parameter from current
solution S picked randonly.
Usage
saa.neighborhood1(f, S, d)
Arguments
f |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
S |
The current solution to find a neighbor |
d |
The distance from current solution S |
Value
The neighbor of solution of S
saa.neighborhoodH
Description
Generates neighbor solutions perturbing half parameters from current
solution S.
Usage
saa.neighborhoodH(f, S, d)
Arguments
f |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
S |
The current solution to find a neighbor |
d |
The distance from current solution S |
Value
The neighbor of solution of S
saa.neighborhoodN
Description
Generates neighbor solutions perturbing all parameters from current
solution S.
Usage
saa.neighborhoodN(f, S, d)
Arguments
f |
An instance of ObjectiveFunction (or subclass) class ObjectiveFunction |
S |
The current solution to find a neighbor |
d |
The distance from current solution S |
Value
The neighbor of solution of S
saa.tbyk
Description
Temperature function t/k
Usage
saa.tbyk(t0, k)
Arguments
t0 |
The current temperature |
k |
The annealing value |
Value
The new temperature
saa.tcte
Description
Temperature function cte * t0
Usage
saa.tcte(t0, k)
Arguments
t0 |
The current temperature |
k |
The annealing value |
Value
The new temperature
saa.texp
Description
Temperature function exponential
Usage
saa.texp(t0, k)
Arguments
t0 |
The current temperature |
k |
The annealing value |
Value
The new temperature
scatterplotlothelper
Description
Simple helper for ploting 3d scaterplots
Usage
scatterplotlothelper(d, x, y, z, title = NULL)
Arguments
d |
A data frame. |
x |
A string with the dataframe column name for x axis |
y |
A string with the dataframe column name for y axis |
z |
A string with the dataframe column name for z axis |
title |
The optional plot title. May be omited. |
Value
A scatter3D plot
searchrow
Description
Search for a value value on a matrix
Usage
searchrow(ddata, value)
Arguments
ddata |
The matrix containing the dataset |
value |
The value to search for |
Value
Boolean TRUE for those indexes matching value
show.comp1
Description
Generates a barplot comparing the number of evalutions for algorithms ("saa","pso","acor","ees1").
Usage
show.comp1(mydata, what, title = NULL)
Arguments
mydata |
The data generated with 'summarize.comp1' |
what |
The name of variable to plot on 'y' axis |
title |
the plot title |
Examples
## Not run:
p.a<- show.comp1(d.cigar4,"evals","(a) Cigar function")
p.b<- show.comp1(d.schaffer4,"evals","(b) Schafer function")
p.c<- show.comp1(d.griewank4,"evals","(c) Griewank function")
p.d<- show.comp1(d.bohachevsky4,"evals","(d) Bohachevsky function")
## End(Not run)
slope
Description
Simple function for calculate the slope on the ith element position
Usage
slope(x, y, i)
Arguments
x |
The x vector |
y |
The y vector |
i |
The position |
Value
The slope
slopes
Description
Calcule all slopes for the discrete x,y series
Usage
slopes(x, y)
Arguments
x |
The x vector |
y |
The y vector |
Value
A vector with all slopes
sortSolution
Description
Sort solution by its respective fitness
Usage
sortSolution(s, f)
Arguments
s |
Problem solution |
f |
The function evaluation for s |
summarize.comp1
Description
Provides as summary with averged values of experimental setup
Usage
summarize.comp1(mydata)
Arguments
mydata |
The data frame generated with 'compare.algorithms1' |
Value
The summarized data
tabu.getNeighbors
Description
create neighbor solutions
Usage
tabu.getNeighbors(tabu, parameters, solution, size)
Arguments
tabu |
The tabu list |
parameters |
The parameter set |
solution |
The current solution |
size |
The neigborhood size |
Value
The neighbor for solution
tabu.istabu
Description
Check whether a solution is present on tabulist
Usage
tabu.istabu(tabulist, solution)
Arguments
tabulist |
The matrix of tabu solutions |
solution |
The solution value to be checked |
Value
Boolean TRUE tabulist contains the solution
upperBound
Description
Checks if parameters is below the upper bounds
Usage
upperBound(particles, factors)
Arguments
particles |
The particle set |
factors |
the defined range for objective function parameters |
Value
The particle inside the valid upper bound
xmeanci1
Description
Calculates confidence interval of mean for provided data with desired confidence level. This functions uses bootstrap resampling scheme for estimanting the CI.
Usage
xmeanci1(x, alpha = 0.95)
Arguments
x |
The data set for which CI will be calculated |
alpha |
The confidence level. The default value is 0.95 (95%) |
Value
The confidence interval for the mean calculated using 'boot.ci'
xmeanci2
Description
Calculates confidence interval of mean for provided data with desired confidence level.
Usage
xmeanci2(x, alpha = 0.95)
Arguments
x |
The data set for which CI will be calculated |
alpha |
The confidence level. The default value is 0.95 (95%) |
Value
The confidence interval for the mean
xyplothelper
Description
Simple helper for ploting xy dispersion points.
Usage
xyplothelper(d, x, y, title = NULL)
Arguments
d |
A data frame. |
x |
A string with the dataframe column name for x axis |
y |
A string with the dataframe column name for y axis |
title |
The optional plot title. May be omited. |
Value
A ggplot2 plot object