Codip: A Hybrid Continuous-variable and Discrete-variable Quantum System for Scalable and Accurate Integer Programming

Codip is a hardware-software co-design framework enabling scalable integer programming on hybrid CV-DV (Continuous-Variable and Discrete-Variable) architectures. By leveraging superconducting cavities for dense integer storage and transmon qubits for non-linear control, Codip bridges the semantic gap between high-level constraints and physical hardware.

This repository contains the reference implementation of the Codip system, including the constraint-preserving intermediate representation (IR), the Hamiltonian-based compiler, and the runtime drift manager.

Key Features

• Hybrid CV-DV Architecture: Implements the "One-mode One-integer" abstraction to reduce memory requirements by 81% compared to standard DV one-hot encoding.
• Constraint-Preserving IR: Analytically computes the kernel of the constraint matrix ($Ax=b$) to construct Hamiltonians that confine evolution strictly within the valid subspace, eliminating the need for penalty terms.
• Drift Manager: A runtime module capable of detecting and mitigating photon-loss errors using subspace-based error mitigation and phase watchdogs.
• High Accuracy: Achieves a $10^2\times$ to $10^5\times$ improvement in success rate compared to state-of-the-art baselines.

Requirements

• Python 3.x
• QuTiP: Used for backend simulation of quantum dynamics and Hamiltonian evolution[cite: 498].
• Other dependencies are listed in pyproject.toml.

Repository Structure

The core source code is located in the cvIP/ directory.

File	Description
cvIP/main.py	Main Entry Point. The primary solver script that orchestrates the compilation and execution flow.
cvIP/integerCons.py	Frontend & IR Translator. Handles variable definitions, extracts constraint equalities, and generates Kernel Hamiltonians.
cvIP/jointMitigate.py	Drift Manager. Implements active monitoring and subspace-based error mitigation (Dual-State Purification).
cvIP/bosonicQAOA.py	Core simulation engine for the CV system QAOA.
cvIP/generate_problems.py	Utility to generate random benchmark subproblems
cvIP/ddqaoa.py	Implementation of Discrete-Variable (DV) baselines for comparison.

Usage

1. Environment Setup

Ensure you have the required dependencies installed. You can install the project in editable mode:

pip install -e .

2. Run the Codip Solver

Use the main script main.py to solve an integer programming instance. [cite_start]This script triggers the compilation and execution workflow[cite: 184].

python cvIP/main.py

(Note: You may need to pass specific arguments such as problem path or layer depth depending on your configuration within main.py.)