generate_integ_pts: Strategy generation

In vpicheny/GPGame: Solving Complex Game Problems using Gaussian Processes

Description

Preprocessing to link strategies and designs.

Usage

generate_integ_pts(
  n.s,
  d,
  nobj,
  x.to.obj = NULL,
  gridtype = "cartesian",
  equilibrium = "NE",
  lb = rep(0, d),
  ub = rep(1, d),
  include.obs = FALSE,
  model = NULL,
  init_set = NULL,
  include_set = NULL,
  seed = 42
)

Arguments

n.s	scalar or vector. If scalar, total number of strategies (to be divided equally among players), otherwise number of strategies per player.
d	number of variables
nobj	number of objectives (or players)
x.to.obj	vector allocating variables to objectives. If not provided, default is 1:nobj, assuming that d=nobj
gridtype	either "cartesian" or "lhs", or a vector to define a different type for each player.
equilibrium	either "NE", "KSE", "CKSE" or "NKSE"
lb, ub	vectors specifying the bounds of the design space, by default [0,1]^d
include.obs	Boolean, if TRUE observations given in model@X are added to the integration points (only for KSE and CKSE)
model	optional list of km models (used if include.obs=TRUE)
init_set	large grid to subsample from
include_set	grid to be included in the larger one generated
seed	random seed used by lhsDesign

Value

A list containing two matrices, integ.pts the design of experiments and expanded.indices the corresponding indices (for NE), and the vector n.s

Examples

##############################################
### 4 variables, 2 players, no filter
##############################################

# Create a 11x8 lattice based on 2 LHS designs
n.s <- c(11,8)
gridtype = "lhs"
# 4D space is split in 2
x.to.obj <- c(1,1,2,2)
integcontrol <- generate_integ_pts(n.s=n.s, d=4, nobj=2, x.to.obj = x.to.obj, gridtype=gridtype)
pairs(integcontrol$integ.pts)

# Create a simple 11x11 grid
integcontrol <- generate_integ_pts(n.s=11^2, d=2, nobj=2, gridtype="cartesian")
pairs(integcontrol$integ.pts)

vpicheny/GPGame documentation built on Jan. 26, 2022, 9:17 a.m.