Skip to main content

A solver for fixed route vehicle charging problems

Project description

frvcpy: An Open-Source Solver for the FRVCP

Fast optimal solutions to rich FRVCPs

What is an FRVCP?

Given an elctric vehicle (EV) that's been assigned some sequence of locations to visit (a fixed route), the fixed route vehicle charging problem (FRVCP) is the problem of finding the optimal insertion of recharging operations into the route that minimize the time required for the EV to traverse that route in an energy-feasible manner.


In a virtual environment with Python 3.6+, frvcpy can be installed via

pip install frvcpy

Testing the installation

import frvcpy.test

Or from the command line:


Using frvcpy

With a compatible instance file (see the schema), solve the FRVCP from a Python script:

from frvcpy import solver

route = [0,40,12,33,38,16,0]        # route to make energy feasible
q_init = 16000                      # vehicle's initial energy level

# using an existing instance from file
frvcp_solver = solver.Solver("instances/frvcpy-instance.json", route, q_init)

# run the algorithm
duration, feas_route = frvcp_solver.solve()

# write solution to file
frvcp_solver.write_solution("my-solution.xml", instance_name="frvcpy-instance")

print(f"Duration: {duration:.4}")
# Duration: 7.339

print(f"Energy-feasible route:\n{feas_route}")
# Energy-feasible route:
# [(0, None), (40, None), (12, None), (33, None), (48, 6673.379615520617), (38, None), (16, None), (0, None)]

Or from the command line:

frvcpy --instance=instances/frvcpy-instance.json --route=0,40,12,33,38,16,0 --qinit=16000 --output=my-solution.xml
# Duration: 7.339
# Energy-feasible route:
# [(0, None), (40, None), (12, None), (33, None), (48, 6673.379615520617), (38, None), (16, None), (0, None)]

Solutions are written in the VRP-REP format for easy importing and visualization with the VRP-REP Mapper (formal solution specification available here).

Note: Example problem instances are available in the instances directory on the project's homepage. For easy access to the example files, we recommend cloning the repository.

Instance Translation

Instance translation is available for some E-VRP instances formatted according to the VRP-REP specification (available here).

Translation can be done with the Python API via

from frvcpy import translator

# Option 1) write the translated instance to file
translator.translate("instances/vrprep-instance.xml", to_filename="instances/my-new-instance.json")

# Option 2) make instance object to be passed directly to the solver
frvcp_instance = translator.translate("instances/vrprep-instance.xml")

Or with the command line:

# from CLI, only option is to write translated instance to file
frvcpy-translate instances/vrprep-instance.xml instances/my-new-instance.json

Note: If an instance ending in ".xml" is passed to the solver, it is assumed to be a VRP-REP instance, and the solver will automatically attempt to translate it.

Translation requirements for VRP-REP instances

frvcpy's translator assumes VRP-REP instances are formatted similarly to the Montoya et al. (2017) instances:

  • CSs are identified as <node> elements having attribute type="2"
  • Charging stations nodes have a <custom> child element which contains a cs_type element
  • For each unique CS type t appearing in those cs_type elements, there exists a charging function element with attribute cs_type=t
  • These function elements are part of a charging_functions element in a vehicle_profile's custom element
  • The depot has node ID 0, the N customers have IDs {1, ..., N}, and the CSs have IDs {N+1, ..., N+C}

A good example of such an instance is the example VRP-REP instance in the repository.

Here is a smaller example meeting these requirements:

<?xml version="1.0" encoding="UTF-8"?>
      <node id="0" type="0">
      <node id="1" type="1">
      <node id="11" type="2">
    <vehicle_profile type="0">
          <function cs_type="fast">
    <request id="1" node="1">

The Solver

To solve FRVCPs, frvcpy implements the labeling algorithm from Froger et al. (2019), providing optimal solutions in low runtime. The algorithm accommodates realistic problem features such as nonlinear charging functions, heterogeneous charging station technologies, and multiple CS visits between stops.

Additional information

For more information about the algorithm used in the solver, see Froger et al. (2019).

A write-up of this package is available on HAL here.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

frvcpy-0.1.1.tar.gz (71.3 kB view hashes)

Uploaded source

Built Distribution

frvcpy-0.1.1-py3-none-any.whl (75.4 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page