Skip to contents

Generate random interaction matrix for GLV model

Source: R/randomA.R
randomA.Rd

Generates a random interaction matrix for Generalized Lotka-Volterra (GLV) model.

Usage

randomA(
  n_species,
  names_species = NULL,
  diagonal = -0.5,
  connectance = 0.2,
  scale_off_diagonal = 0.1,
  mutualism = 1,
  commensalism = 1,
  parasitism = 1,
  amensalism = 1,
  competition = 1,
  interactions = NULL,
  symmetric = FALSE,
  list_A = NULL
)

Arguments

n_species

Integer: number of species

names_species

Character: names of species. If NULL, paste0("sp", seq_len(n_species)) is used. (default: names_species = NULL)

diagonal

Values defining the strength of self-interactions. Input can be a number (will be applied to all species) or a vector of length n_species. Positive self-interaction values lead to exponential growth. (default: diagonal = -0.5)

connectance

Numeric frequency of inter-species interactions. i.e. proportion of non-zero off-diagonal terms. Should be in the interval 0 <= connectance <= 1. (default: connectance = 0.2)

scale_off_diagonal

Numeric: scale of the off-diagonal elements compared to the diagonal. (default: scale_off_diagonal = 0.1)

mutualism

Numeric: relative proportion of interactions terms consistent with mutualism (positive <-> positive) (default: mutualism = 1)

commensalism

Numeric: relative proportion of interactions terms consistent with commensalism (positive <-> neutral) (default: commensalism = 1)

parasitism

Numeric: relative proportion of interactions terms consistent with parasitism (positive <-> negative) (default: parasitism = 1)

amensalism

Numeric: relative proportion of interactions terms consistent with amensalism (neutral <-> negative) (default: amensalism = 1)

competition

Numeric: relative proportion of interactions terms consistent with competition (negative <-> negative) (default: competition = 1)

interactions

Numeric: values of the n_species^2 pairwise interaction strengths. Diagonal terms will be replaced by the 'diagonal' parameter If NULL, interactions are drawn from runif(n_species^2, min=0, max=abs(diagonal)). Negative values are first converted to positive then the signs are defined by the relative weights of the biological interactions (i.e. mutualism, commensalism, parasitism, amensalism, competition) (default: interactions = NULL)

symmetric

Logical: whether the strength of mutualistic and competitive interactions are symmetric. This is implemented by overwrite a half of the matrix, so the proportions of interactions might deviate from expectations. (default: symmetric=FALSE)

list_A

List: a list of matrices generated by randomA. Used to support different groups of interactions. If NULL (by default), no group is considered. Otherwise the given list of matrices will overwrite values around the diagonal. (default: list_A = NULL)

Value

randomA returns a matrix A with dimensions (n_species x n_species)

Examples


dense_A <- randomA(
    n_species = 10,
    scale_off_diagonal = 1,
    diagonal = -1.0,
    connectance = 0.9
)

sparse_A <- randomA(
    n_species = 10,
    diagonal = -1.0,
    connectance = 0.09
)

user_interactions <- rbeta(n = 10^2, .5, .5)
user_A <- randomA(n_species = 10, interactions = user_interactions)

competitive_A <- randomA(
    n_species = 10,
    mutualism = 0,
    commensalism = 0,
    parasitism = 0,
    amensalism = 0,
    competition = 1,
    connectance = 1,
    scale_off_diagonal = 1
)

parasitism_A <- randomA(
    n_species = 10,
    mutualism = 0,
    commensalism = 0,
    parasitism = 1,
    amensalism = 0,
    competition = 0,
    connectance = 1,
    scale_off_diagonal = 1,
    symmetric = TRUE
)

list_A <- list(dense_A, sparse_A, competitive_A, parasitism_A)
groupA <- randomA(n_species = 40, list_A = list_A)