RSSched MATSim Client

Convert MATSim simulation outputs to rssched-solver requests. Part of the Innosuisse project Rolling Stock Scheduling (RSSched) by SBB and ETH Zurich.

The client applies the Pipes and Filters architectural pattern to process the source data in a sequential manner. The pattern structures a system as a series of processing elements (filters) connected by channels (pipes). Each filter performs a specific processing task on the data it receives and passes the processed data to the next filter through the pipes.

Project Structure

The project has the following package structure:

• ch.sbb.rssched.client: Client to bind to the service and send requests.
  • config: Configure the request and the pipeline to build it.
    • selection: Define strategies to filter the transit lines to be considered.
  • dto: implements interface to the rolling stock scheduling solver of the ETH Zurich.
    • request: Build request for solver.
    • response: Parse solver response.
  • pipeline``:
    • core: Provides generic classes for implementing the Pipes and Filters architectural pattern, such as the data source, pipe (=data transport), filter, pipeline, and result sink.
    • passenger: Passenger data analysis pipeline.
    • scenario: Scenario processing pipeline, includes various filters and data manipulations, such as scenario masking or attribute removal.
    • request: Create a solver request, by collecting the results from the passenger and scenario pipelines.
    • response: Send the request to the solver and await response.
    • utils: Common helper and utility classes.

Usage

To get started with the whole RSSched project, have a look at this step-by-step instruction.

Using Command-Line Application

Read the request configuration from an Excel file and optionally send the request to the solver using the RsschedMatsimClient:

mvn exec:java -Dexec.args="path/to/config_file.xlsx -h localhost -p 3000 -d"

Options:

• Required:
  • config_file: Path to the Excel configuration file.
• Optional:
  • -h / --host: Scheduler base URL (default: "http://localhost").
  • -p / --port: Scheduler port (default: 3000).
  • -d / --dry-run: If present, do not send the request to the solver (default: false).

See kelheim-v3.0-25pct.rssched_request_config.xlsx for reference of a request configuration.

Note: To be able to run the integration test using the command line app, first execute the integration tests, which downloads the needed matsim run outputs, see the paragraph Testing below.

Using Client in Java

Set up the RsschedRequestClient and send a request to the rolling stock scheduling solver service using the RsschedRequestConfig to configure it. Optionally implement a transit line filter strategy.

import ch.sbb.rssched.client.RsschedMatsimClient;
import ch.sbb.rssched.client.config.RsschedRequestConfig;
import ch.sbb.rssched.client.dto.response.Response;

public class RunExample {

    private static final String SCHEDULER_BASE_URL = "http://localhost";
    private static final int SCHEDULER_PORT = 3000;

    public static void main(String[] args) {
        RsschedRequestConfig config = RsschedRequestConfig.builder()
                .setInputDirectory("path/to/input/directory")
                .setOutputDirectory("path/to/output/directory")
                .setRunId("runId")
                .setInstanceId("rss001")
                // optionally set transit line filter, default is no filtering
                .setFilterStrategy(scenario -> {
                    // implementation...
                    return null;
                }).buildWithDefaults();

        // optionally set sample size factor, default is 1.0
        config.getGlobal().setSampleSize(0.25);

        RsschedMatsimClient client = new RsschedMatsimClient(SCHEDULER_BASE_URL, SCHEDULER_PORT);
        Response response = client.process(config);
    }

}

Alternatively use an Excel file to configure the request to the solver:

import ch.sbb.rssched.client.RsschedMatsimClient;
import ch.sbb.rssched.client.config.RsschedRequestConfig;
import ch.sbb.rssched.client.config.RsschedRequestConfigReader;
import ch.sbb.rssched.client.dto.response.Response;

import java.io.IOException;

public class RunExample {

    private static final String REQUEST_CONFIG_XLSX = "rssched_request_config.xlsx";
    private static final String SCHEDULER_BASE_URL = "http://localhost";
    private static final int SCHEDULER_PORT = 3000;

    public static void main(String[] args) throws IOException {
        RsschedRequestConfig config = new RsschedRequestConfigReader().readExcelFile(REQUEST_CONFIG_XLSX);
        RsschedMatsimClient client = new RsschedMatsimClient(SCHEDULER_BASE_URL, SCHEDULER_PORT);
        Response response = client.process(config);
    }

}

Note: There is a hard limit of 500 locations per instance, since the deadhead trip matrix grows exponentially.

Testing

Run the unit tests:

mvn test

Run the integration test to see the pipeline in action (needs a running solver on localhost and port 3000):

mvn verify -Dit.test=RsschedMatsimClientIT

---

© 2024 SBB CFF FFS. Licensed under GPL-3.0.