Creating an instance programmatically

PathWyse can build and solve a problem entirely from Python, without an instance file. The full example is at examples/python_application/pathwyse_instance_creation.py. The walkthrough below creates a small 5-node RCSPP with a capacity resource.

Step 1) Import PathWyse wrapper and create a solver

from bin.wrapper import PWSolver
pw = PWSolver()

Step 2) Create a problem and set its basic properties

pw.createProblem()
pw.setProblemName("Toy")
pw.setNumNodes(5)
pw.setDirected(True)
pw.setCyclic(True)
pw.setOrigin(0)
pw.setDestination(4)

Step 3) Initialize internal data structures

pw.initProblem()

This must be called before adding resources.

Step 4) Add resources and set their global bounds

cap = pw.addResource("CAP")
pw.setResBounds(cap, 0, 12)

Step 5) Add arcs and their costs

arc_costs = [
    (0, 1, 54000), (0, 2, 35000), (0, 3, 54000),
    (1, 0, 54000), (1, 2, 87000), (1, 3, 75000), (1, 4, 79000),
    (2, 0, 35000), (2, 1, 87000), (2, 3, 54000),
    (3, 0, 54000), (3, 1, 75000), (3, 2, 54000), (3, 4, 98000),
    (4, 0, 45000), (4, 1, 79000), (4, 2, 59000), (4, 3, 98000),
]
for i, j, c in arc_costs:
    pw.addArc(i, j)
    pw.setArcCost(i, j, c)

Step 6) Set node costs

node_costs = {0: -7435, 1: 0, 2: -200000, 3: -7839, 4: -4000}
for node, c in node_costs.items():
    pw.setNodeCost(node, c)

Step 7) Set node resource consumptions

node_consumption = {0: 0, 1: 2, 2: 1, 3: 6, 4: 6}
for node, c in node_consumption.items():
    pw.setResNodeConsumption(cap, node, c)

Step 8) Build the problem

pw.buildProblem()

This finalizes the problem (scales costs, validates structure) and must be called after all data has been set.

Step 9) Setup algorithms, solve, and print

pw.setupAlgorithms()
pw.solve()
pw.printBestSolution()