XpressMP Java Examples

2025-12-08 · Source: OR in an OB World · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Intermediate, quick

Summary

FICO XpressMP's Java API is undergoing a transition from an older "thin wrapper" (XPRSprob class) to a newer, more object-oriented version (XpressProblem class). This transition explains the presence of seemingly duplicate examples within the XpressMP installation, specifically in the `examples/solver/optimizer/java` directory and its `objects` subfolder. Examples in the `objects` folder utilize the recommended new API, while those in the parent directory use the older API. A key distinction between the two APIs, particularly evident in examples involving callbacks like the Traveling Salesperson Problem (TSP), lies in how they handle model modifications during the presolve stage. The older API allows direct interaction with the presolved model, requiring manual "crushing" of cuts, whereas the newer API abstracts this, allowing users to define cuts in terms of original variables, with Xpress handling the conversion.

Key takeaway

For Java developers working with FICO XpressMP, you should prioritize using the newer, object-oriented API based on the XpressProblem class. This approach simplifies handling model transformations and cuts during the presolve stage, as Xpress manages the "crushing" process automatically. Refer to the examples located in the `objects` subfolder within the XpressMP installation for guidance on implementing solutions with the recommended API, ensuring your code benefits from the streamlined design.

Key insights

FICO XpressMP's Java API is transitioning to an object-oriented design, simplifying cut management during presolve.

Principles

• Newer APIs often abstract complexity.
• Object-oriented APIs can simplify development.

In practice

• Use the XpressProblem class for new Java development.
• Consult `objects` folder examples for new API usage.

Topics

• FICO XpressMP
• Java API
• API Transition
• Optimization Callbacks
• Presolve Stage

Best for: Software Engineer, Machine Learning Engineer, AI Engineer

Related on AIssential

• Benders Decomposition in XpressMP — Implementing Benders decomposition in XpressMP via Java callbacks presents unique challenges and workarounds.
• Predicting Performance of Symbolic and Prompt Programs with Examples — Symbolic programs are "all-or-nothing" in performance, while prompt programs are "diffuse" and less predictable.
• FastAPI for MLOps: Python Project Structure and API Best Practices — Robust software engineering, not just models, drives successful, stable, and scalable MLOps systems.
• Parax v0.5: Parametric Modeling in JAX [P] — Parax v0.5 offers flexible, opt-in parametric modeling tools for JAX with an extensible API.
• Build your own AI tools to fill specific workflow gaps — Custom AI tools can bridge workflow gaps by enabling AI models to interpret and generate structured visual designs.
• LLM-Pruning Collection: A JAX Based Repo For Structured And Unstructured LLM Compression — LLM-Pruning Collection unifies diverse LLM pruning methods within a consistent JAX-based framework for reproducible comparison.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by OR in an OB World.