anytime pmnk-landscapes

Dependencies

CLI11

Compilation

cmake -B build-S . -DCMAKE_BUILD_TYPE=Release
cmake --build build

Usage

General

Usage: anytime-pmnk-landscapes [OPTIONS] instance ALGORITHM

Positionals:
  instance .dat:FILE REQUIRED           
      = pmnk-landscapes instance file path
      (instances can be generated using the rmnkGenerator.R script).

Options:
  -m,--maxeval UINT:NONNEGATIVE REQUIRED Needs: instance
            = maximum number of evaluations to be performed (stopping criterion).
  -s,--seed UINT:NONNEGATIVE Needs: instance
            = pseudo random generator seed used by the search heuristics.
  -o,--output TEXT Needs: instance      
            = specifies the file to which the output stream should be redirected.
  -r,--hvref FLOAT ... Needs: instance  
            = reference point considered in the hypervolume calculation.
  -h,--help                             
            = print this help message and exit.
  -H,--help-all                         
            = expand all help.

Algorithms:
  GSEMO    Run the global simple evolutionary multiobjective optimizer algorithm
           on the instance.
  PLS      Run the pareto local search algorithm on the instance.
  IBEA     Run the indicator-based evolutionary algorithm on the instance.

Examples

• Common Usage

./anytime-pmnk-landscapes

GSEMO

Run the global simple evolutionary multiobjective optimizer algorithm
on the instance.
Usage: anytime-pmnk-landscapes [OPTIONS] instance GSEMO

Examples

PLS

Run the pareto local search algorithm on the instance.
Usage: anytime-pmnk-landscapes [OPTIONS] instance PLS [OPTIONS]

Options:
  -h,--help       
      = print this help message and exit.
  -H,--help-all                         
      = expand all help.
  -a,--pls-acceptance-criterion 
    ENUM:value in {BOTH->2,DOMINATING->1,NON_DOMINATING->0} OR {2,1,0}
        = acceptance criterion considered whilst running pls.
            => (NON_DOMINATING): accept every non-dominated neighbor.
            => (DOMINATING): accept only neighbors that dominate current solution.
            => (BOTH): first try to accept only neighbors that dominate
                the current solution, if none exist accept non-dominated solutions.
  -e,--pls-neighborhood-exploration 
    ENUM:value in {BEST_IMPROVEMENT->0,BOTH->2,FIRST_IMPROVEMENT->1} OR {0,2,1}
        = neighborhood exploration criterion considered whilst running pls.
            => (BEST_IMPROVEMENT): explore every acceptable neighboor.
            => (FIRST_IMPROVEMENT): stop once on neighbor is accepted.
            => (BOTH): use FIRST_IMPROVEMENT until PLS stops, afterwards use BEST_IMPROVEMENT

IBEA

Run the indicator-based evolutionary algorithm on the instance.
Usage: anytime-pmnk-landscapes IBEA [OPTIONS] INDICATOR MUTATION CROSSOVER SELECTION

Options:
  -h,--help   
    = print this help message and exit.
  -H,--help-all                         
    = expand all help.
  -p,--pop-size UINT:NONNEGATIVE REQUIRED
    = max population size.
  -g,--generations UINT:NONNEGATIVE REQUIRED
    = number of generations. (stopping criterion)
  -k,--scaling-factor FLOAT:NONNEGATIVE REQUIRED
    = scaling factor.
  -a,--adaptive                         
    = use the adaptive version of the algorithm

Indicators:
IHD
  Run using the hypervolume indicator

EPS
  Run using the epsilon (+) indicator

Mutation Operators:
UniformMutation
  Run using a uniform mutation operator
  Options:
    -p,--mutation-probability FLOAT:FLOAT in [0 - 1] REQUIRED
        = probability of occurrence of a mutation in the individual's genotype

Crossover Operators:
NPointCrossover
  Run using a n-point crossover operator
  Options:
    -p,--crossover_probability FLOAT:FLOAT in [0 - 1] REQUIRED
        = probability of occurrence of a mutation in the individual's genotype
    -n,--n-points UINT:NONNEGATIVE REQUIRED
        = number of randomly picked crossover points

UniformCrossover
  Run using a uniform crossover operator
  Options:
    -p,--crossover_probability FLOAT:FLOAT in [0 - 1] REQUIRED
        = probability of occurrence of a mutation in the individual's genotype


Selection Operators:
KWayTournament
  Run using a k-way tournament selection operator
  Options:
    -s,--matting-pool-size UINT:NONNEGATIVE REQUIRED
        = target size of the matting pool to be obtained from the selection step
    -t,--tournament-size UINT:NONNEGATIVE REQUIRED
        = size of the tournament used for individual selection

Examples

API

For more information about the usage of this api please consult the available documentation

Acknowledgments

This project was only made possible thanks to the support of a research grant at the Center for Informatics and Systems of the University of Coimbra (CISUC), with reference number UIDB/00326/2020, awarded by the Foundation for Science and Technology (FCT).

Contributors

• Alexandre D. Jesus