skim.paths: Skim a network
In travelr: Functions for Travel Demand Modeling

Description Usage Arguments Details Value Author(s) References Examples

Apply a function to link values on each path

1	skim.paths(paths,costs,empty.val=0.0,FUN=sum,...)

`paths`	A `path.set` build by `build.paths`
`costs`	A vector or matrix of costs for the network
`empty.val`	A value to store in skim matrix cells that are not crossed by any path (e.g. intrazonal cells)
`FUN`	The function to apply to the `costs` vector associated with each of the `paths`; `FUN` should return a scalar value that can be stored in a numeric matrix
`...`	Additional arguments to be passed to `FUN`

“Skimming a network” means computing a value for the shortest path between each origin and each destination in the network. skim.paths returns a square matrix where the number of rows and columns corresponds to the number of zones in the network, and the value in each cell is the result of applying FUN to the costs values corresponding to the links on each shortest path.

Most commonly, skimming is used to add up a cost value for all the links on the paths between all the origin/destination pairs. It is also common to add up the values of other attributes (total distance, congested or free-flow time, miles of interstate highway, etc.).

Sometimes, there are origin and destination pairs that do not have a path available. Most commonly, these will be intrazonal cells (the diagonal of the resulting matrix, where the origin and destination are the same). Otherwise, they will indicate that the origin and destination are not connected, which can be useful to know when debugging a highway network.

The following functionality is not implemented yet: However, by applying a different function, skimming can extract a wide range of values, including the mean, minimum, or maximum link cost on a path, or perhaps even a true or false value depending on whether a certain link exists in a path (which yields the same result as .intercept.paths, only more slowly). In general, if this function were named to appeal to people who are primarily R users, rather than people who are primarily travel modelers, it might also be called "path.apply" since the basic operation works very much like tapply applied to a list of vectors of varying lengths.

skim.paths returns a numeric origin/destination matrix where each cell contains the corresponding skim result; see details above.

Jeremy Raw

Martin, W.A., and McGuckin, N.A., 1998, NCHRP Report 365: Travel Estimation Techniques for Urban Planning, National Cooperative Highway Research Program / Transportation Research Board

# A tiny, complete travel model using travelr functions
data(SiouxFalls)

# Trip Generation
productions<-rowSums(SiouxFalls.od)
attractions<-colSums(SiouxFalls.od)

# Highway Skims
cost.function<-with(SiouxFalls.net$Links,function(...)FFTime)
aclass <- make.assignment.class(SiouxFalls.net,"All",SiouxFalls.od)
aset <- new.assignment.set(SiouxFalls.net,list(All=aclass),cost.volume.type="vector",cost.function=cost.function)
paths <- build.paths(aset,aset$ff.cost)
travel.times <- skim.paths(paths,aset$ff.cost)[["All"]] # only one purpose: "All trips"

# Trip Distribution (Gravity Model with gamma function)
base.distribution <- hwy.gamma.function(travel.times,-0.02,-0.123) # HBW coefficients from NCHRP 365
trip.table <- ipf(base.distribution,list(rows=productions, cols=attractions),method="absolute")
print(round(trip.table,2))
aset <- hwy.update.demand(aset,"All",trip.table)

# Trip Assignment
assignment.results <- highway.assign(aset,method="Frank.Wolfe")
loaded.links <- assignment.results$volumes