igd: IGD: Inverted Generational Distance

Description Usage Arguments Value References Examples

View source: R/ps_metrics.R

Description

The IGD is a perfomance measure function of Pareto front fidelity and corresponds to the average distance between all designs in the true set and the closest design of the current set. Thus, the lower the IGD value, the better the front is.

Usage

1
igd(aps, tps, method = "manhattan", norm = TRUE)

Arguments

aps

An object of type ps containing the "actual" pareto front

tps

An object of type ps containing the "true" pareto front

method

String stating which distance measure to be used. This must be one of: "euclidean" or "manhattan" (default).

norm

Logical (default: TRUE) indicating if both fronts should be normalized.

Value

returns the IGD metric

References

Shimoyama, K., Jeong, S., & Obayashi, S. (2013, June). Kriging-surrogate-based optimization considering expected hypervolume improvement in non-constrained many-objective test problems. In 2013 IEEE Congress on Evolutionary Computation (pp. 658-665). IEEE.

Examples

1
2
3
4
5
6
7
aps <- ps(matrix(rnorm(1:1000),ncol=2))
tps <- ps(matrix(rnorm(1:2000),ncol=2))
igd(aps,tps)

tps <-nowacki_beam_tps$set[1:50 * 10,]
aps <- tps * 1.2
igd(aps,tps)

coldfir3/moko documentation built on May 13, 2019, 8:49 p.m.