org.graphstream.algorithm.APSP$APSPInfo.getShortestPathTo 处的 NullPointerException - 图形问题

Question

我正在开发一种算法来寻找两个机场之间的最佳路线。我遇到了All Pairs Shortest Path (APSP) 和GraphStream 库的概念。

内容是从文本文件中读取的。每条线代表一个航空公司中途停留：

PEK,LAX,5
ATL,HND,75
ATL,LAX,20
DXB,HND,5
ATL,PEK,10
PEK,HND,40
LAX,DXB,20
ATL,DXB,55

因为是涉及机票的问题，所以我将第 3 列称为“价格”，但是当我尝试在其中输入实际价格值时，一些路线停止工作。此外，其中一项功能是向文件添加新行。我在为新的中途停留添加任何我想要的值时遇到问题。

阅读a StackExchange's Math forum thread，在创建图形字符串之前，我尝试将所有值除以最大权重，但没有成功：

String graphString;
String graphStringHeader = "DGS004\nmy 0 0\n";
String graphStringNodes = "";
String graphStringEdges = "";

BigDecimal biggestPrice = BigDecimal.ZERO;

while (stLines.hasMoreTokens()) {
    line = stLines.nextToken(System.getProperty("line.separator"));

    final StringTokenizer stColumns = new StringTokenizer(line);

    Airport originAux = null;
    Airport destinationAux = null;
    BigDecimal priceAux = null;

    String column;
    Integer columnIndex = 1;

    while (stColumns.hasMoreTokens()) {
        column = stColumns.nextToken(",");

        if (columnIndex == 1) {
            originAux = new Airport(column);

            if (!nodes.contains(column)) {
                // an = add node
                graphStringNodes += "an " + column + " \n";

                nodes.add(column);
            }
        } else if (columnIndex == 2) {
            destinationAux = new Airport(column);

            if (!nodes.contains(column)) {
                // an = add node
                graphStringNodes += "an " + column + " \n";

                nodes.add(column);
            }
        } else if (columnIndex == 3) {
            double parsedPreco = Double.parseDouble(column);
            priceAux = BigDecimal.valueOf(parsedPreco);

            /**
             * Normalizing values.
             */
            if (priceAux.intValue() > biggestPrice.intValue()) {
                biggestPrice = priceAux;
            } else {
                priceAux = priceAux.divide(biggestPrice, 2, RoundingMode.HALF_UP);
            }

            edges.add(originAux.getName() + "," + destinationAux.getName()+ "," + priceAux.intValue());

            stopovers.add(new Stopover(originAux, destinationAux, priceAux));

            // ae = add edge
            graphStringEdges += "ae " + originAux.getName() + "_" + destinationAux.getName() + " "
                    + originAux.getName() + " > " + destinationAux.getName()
                    + " price:" + priceAux.intValue()
                    + " \n";
        }

        columnIndex++;
    }

    lineIndex++;
}

graphString = graphStringHeader + graphStringNodes + graphStringEdges;

这是抛出的异常：

java.lang.NullPointerException: null
    at org.graphstream.graph.Path.add(Path.java:230)
    at org.graphstream.algorithm.APSP$APSPInfo.getShortestPathTo(APSP.java:594)
    at service.RouteFinderService.findRoute(RouteFinderService.java:183)

服务的第 230 行是：Path shortestPath = edge.getShortestPathTo(destination.getName())。参数为目的地机场代码。

我如何实例化 Graph 和 APSP 对象：

Graph graph = new DefaultGraph("BestRouteGraph");
ByteArrayInputStream bs = new ByteArrayInputStream(graphString.getBytes());

FileSourceDGS source = new FileSourceDGS();
source.addSink(graph);

source.readAll(bs);

APSP apsp = new APSP();
apsp.init(graph);
apsp.setDirected(false);
apsp.setWeightAttributeName("price");

apsp.compute();

Answer 1

at org.graphstream.graph.Path.add(Path.java:230)

指APSPInfo类中getOpposite方法的使用

nodePath.push(edge.getOpposite(from));

确保检查节点是否存在路径：

for(Node n : graph.getNodeSet()) {          
    for(Node n2 : graph.getNodeSet()) {
        if (n != n2) {
            System.out.println("Shortest Path between "+n+" and "+n2+" :");

            try {
                APSPInfo nodeInfo = n.getAttribute("APSPInfo");
                System.out.println(nodeInfo.getShortestPathTo(n2.getId()));
            } catch (Exception e) {
                System.out.println("no path");
            }
        }
    }
}

例如，在您的数据集中，没有如图所示的 ATL 路径。

这是你应该得到的结果：

Shortest Path between PEK and LAX :
[PEK, LAX]
Shortest Path between PEK and ATL :
no path
Shortest Path between PEK and HND :
[PEK, LAX, DXB, HND]
Shortest Path between PEK and DXB :
[PEK, LAX, DXB]
Shortest Path between LAX and PEK :
no path
Shortest Path between LAX and ATL :
no path
Shortest Path between LAX and HND :
[LAX, DXB, HND]
Shortest Path between LAX and DXB :
[LAX, DXB]
Shortest Path between ATL and PEK :
[ATL, PEK]
Shortest Path between ATL and LAX :
[ATL, PEK, LAX]
Shortest Path between ATL and HND :
[ATL, PEK, LAX, DXB, HND]
Shortest Path between ATL and DXB :
[ATL, PEK, LAX, DXB]
Shortest Path between HND and PEK :
no path
Shortest Path between HND and LAX :
no path
Shortest Path between HND and ATL :
no path
Shortest Path between HND and DXB :
no path
Shortest Path between DXB and PEK :
no path
Shortest Path between DXB and LAX :
no path
Shortest Path between DXB and ATL :
no path
Shortest Path between DXB and HND :
[DXB, HND]