Public

getWeighting

<init>

doWork

Calculates the priority of adjNode v without changing the graph. Warning: the
calculated priority mu

calculatePriority

close

contractNodes

getTimesAsString

initFromGraph

prepareNodes

Introduces the necessary shortcuts for adjNode v in the graph.

addShortcuts

Finds shortcuts, does not change the underlying graph.

findShortcuts

fromGraphHopperStorage

getCoords

getOrigEdgeCount

setLazyUpdates

Specifies how often a log message should be printed. Specify something around 20
(20% of the start n

setLogMessages

setNeighborUpdates

setOrigEdgeCount

The higher the values are the longer the preparation takes but the less
shortcuts are produced.

setPeriodicUpdates

createAStarBidirection

FileOutputStream

Creating JSON documents from java classes using gson

A flow layout arranges components in a left-to-right flow, much like lines of
text in a paragraph. F

java.awt.FlowLayout

An InputStreamReader is a bridge from byte streams to character streams: It
reads bytes and decodes 

java.io.InputStreamReader

An immutable arbitrary-precision signed decimal.A value is represented by an
arbitrary-precision "un

java.math.BigDecimal

Abstract class for representing access to a system resource. All permissions
have a name (whose inte

java.security.Permission

 Calendar is an abstract base class for converting between a Date object and a
set of integer fields

java.util.Calendar

javax.naming.Reference

scheduleAtFixedRate

setContentView

getSharedPreferences

/**
     * Calculates the priority of adjNode v without changing the graph. Warning: the calculated
     * priority must NOT depend on priority(v) and therefor findShortcuts should also not depend on
     * the priority(v). Otherwise updating the priority before contracting in contractNodes() could
     * lead to a slowish or even endless loop.
     */

// set of shortcuts that would be added if adjNode v would be contracted next.

//        System.out.println(v + "\t " + tmpShortcuts);

// # huge influence: the bigger the less shortcuts gets created and the faster is the preparation

// every adjNode has an 'original edge' number associated. initially it is r=1

// when a new shortcut is introduced then r of the associated edges is summed up:

// r(u,w)=r(u,v)+r(v,w) now we can define

// originalEdgesCount = σ(v) := sum_{ (u,w) ∈ shortcuts(v) } of r(u, w)

//        for (Shortcut sc : tmpShortcuts) {

//            originalEdgesCount += sc.originalEdges;

// # lowest influence on preparation speed or shortcut creation count

// (but according to paper should speed up queries)

// number of already contracted neighbors of v

// from shortcuts we can compute the edgeDifference

// # low influence: with it the shortcut creation is slightly faster

// |shortcuts(v)| − |{(u, v) | v uncontracted}| − |{(v, w) | v uncontracted}|

// meanDegree is used instead of outDegree+inDegree as if one adjNode is in both directions

// only one bucket memory is used. Additionally one shortcut could also stand for two directions.

// according to the paper do a simple linear combination of the properties to get the priority.

// this is the current optimum for unterfranken:

// # lowest influence on preparation speed or shortcut creation count 

// |shortcuts(v)| − |{(u, v) | v uncontracted}| − |{(v, w) | v uncontracted}|        

/**
     * Introduces the necessary shortcuts for adjNode v in the graph.
     */

// check if we need to update some existing shortcut in the graph

// note: flags overwrite weight => call first

// preparation takes longer but queries are slightly faster with preparation

// disable lazy updates for last x percentage of nodes as preparation is then a lot slower

// and query time does not really benefit

// according to paper "Polynomial-time Construction of Contraction Hierarchies for Multi-criteria Objectives" by Funke and Storandt

// we don't need to wait for all nodes to be contracted

// Recompute priority of uncontracted neighbors.

// Without neighbor updates preparation is faster but we need them

// to slightly improve query time. Also if not applied too often it decreases the shortcut number.

// periodically update priorities of ALL nodes

"Cannot prepare as no unprepared nodes where found. Called preparation twice?"

// current node got more important => insert as new value and contract it later

// skipped nodes are already set to maxLevel

// Preparation works only once so we can release temporary data.

// The preparation object itself has to be intact to create the algorithm.

How to use findShortcutsmethodin com.graphhopper.routing.ch.PrepareContractionHierarchies

Best Java code snippets using com.graphhopper.routing.ch.PrepareContractionHierarchies.findShortcuts (Showing top 6 results out of 315)

How to use
findShortcuts
method
in
com.graphhopper.routing.ch.PrepareContractionHierarchies