# Plotting bi-objective models with two variables In gMOIP: Tools for 2D and 3D Plots of Single and Multi-Objective Linear/Integer Programming Models

```library(knitr)
opts_chunk\$set(
collapse = TRUE,
#cache = TRUE, autodep = TRUE,
comment = "#>",
fig.show = "asis",
warning=FALSE, message=FALSE, include = TRUE,
out.width = "99%", fig.width = 8, fig.align = "center", fig.asp = 0.62
)

if (!requireNamespace("rmarkdown", quietly = TRUE) || !rmarkdown::pandoc_available("1.14")) {
warning(call. = FALSE, "These vignettes assume rmarkdown and pandoc version 1.14 (or higher). These were not found. Older versions will not work.")
knitr::knit_exit()
}
```

With `gMOIP` you can make plots of the criterion space for bi-objective models (linear programming (LP), integer linear programming (ILP), or mixed integer linear programming (MILP)). This vignette gives examples on how to make plots of both the solution and criterion space.

```library(gMOIP)
```

We define a function for grouping plots of the solution and criterion space:

```plotBiObj2D <- function(A, b, obj,
type = rep("c", ncol(A)),
crit = "max",
faces = rep("c", ncol(A)),
plotFaces = TRUE,
plotFeasible = TRUE,
plotOptimum = FALSE,
labels = "numb",
{
p1 <- plotPolytope(A, b, type = type, crit = crit, faces = faces, plotFaces = plotFaces,
plotFeasible = plotFeasible, plotOptimum = plotOptimum, labels = labels) +
ggplot2::ggtitle("Solution space")
p2 <- plotCriterion2D(A, b, obj, type = type, crit = crit, addTriangles = addTriangles,
ggplot2::ggtitle("Criterion space")
gridExtra::grid.arrange(p1, p2, nrow = 1)
}
```

Let us define the constraints:

```A <- matrix(c(-3,2,2,4,9,10), ncol = 2, byrow = TRUE)
b <- c(3,27,90)
```

First let us have a look at a LP model:

```obj <- matrix(
c(7, -10, # first criterion
-10, -10), # second criterion
nrow = 2)
plotBiObj2D(A, b, obj, addTriangles = FALSE)
```

Note the non-dominated (Pareto) set consists of all supported extreme non-dominated points (illustrated with triangles) and the line segments between them.

ILP models with different criteria (maximize):

```obj <- matrix(c(7, -10, -10, -10), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)))
obj <- matrix(c(3, -1, -2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)))
obj <- matrix(c(-7, -1, -5, 5), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)))
obj <- matrix(c(-1, -1, 2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)))
```

Note the non-dominated set consists of all points in black (with shape supported extreme:triangle, supported non-extreme:round, unsupported:round (not on the border)). The triangles drawn using the extreme non-dominated points illustrate areas where unsupported non-dominated points may be found. A point in the solution space is identified in the criterion space using the same number.

ILP models with different criteria (minimize):

```obj <- matrix(c(7, -10, -10, -10), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)), crit = "min")
obj <- matrix(c(3, -1, -2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)), crit = "min")
obj <- matrix(c(-7, -1, -5, 5), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)), crit = "min")
obj <- matrix(c(-1, -1, 2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = rep("i", ncol(A)), crit = "min")
```

MILP model (\$x_1\$ integer) with different criteria (maximize):

```obj <- matrix(c(7, -10, -10, -10), nrow = 2)
plotBiObj2D(A, b, obj, type = c("i", "c"))
obj <- matrix(c(3, -1, -2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = c("i", "c"))
obj <- matrix(c(-7, -1, -5, 5), nrow = 2)
plotBiObj2D(A, b, obj, type = c("i", "c"))
obj <- matrix(c(-1, -1, 2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = c("i", "c"))
```

Note the solution space now consists to segments and hence the non-dominated set may consist of points and segments (open and closed). Note these segments are not highlighted in the current version of `gMOIP`.

MILP model (\$x_2\$ integer) with different criteria (minimize):

```##
obj <- matrix(c(7, -10, -10, -10), nrow = 2)
plotBiObj2D(A, b, obj, type = c("c", "i"), crit = "min")
obj <- matrix(c(3, -1, -2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = c("c", "i"), crit = "min")
obj <- matrix(c(-7, -1, -5, 5), nrow = 2)
plotBiObj2D(A, b, obj, type = c("c", "i"), crit = "min")
obj <- matrix(c(-1, -1, 2, 2), nrow = 2)
plotBiObj2D(A, b, obj, type = c("c", "i"), crit = "min")
```

## Saving plots to LaTeX

If you write a paper using LaTeX, you may create a TikZ file of the plot for LaTeX using

```library(tikzDevice)
tikz(file = "plot_polytope.tex", standAlone=F, width = 7, height = 6)
plotPolytope(
A,
b,
obj,
type = rep("i", ncol(A)),
crit = "max",
faces = rep("c", ncol(A)),
plotFaces = TRUE,
plotFeasible = TRUE,
plotOptimum = TRUE,
labels = "n",
latex = TRUE
)
dev.off()
```

## Try the gMOIP package in your browser

Any scripts or data that you put into this service are public.

gMOIP documentation built on Aug. 23, 2021, 5:09 p.m.