LSDO Lab
Software

ALBCD

Provably convergent distributed optimization.

Version 1.0

Distribution Statement A. Approved for public release: distribution is unlimited. Approved AFRL-2026-1671 28-09-2026.

Source code Documentation

ALBCD solves distributed multidisciplinary design optimization problems by pairing an augmented Lagrangian outer loop with block coordinate descent over subproblems. The outer loop enforces coupling constraints through multiplier and penalty updates, and with a proximal term every limit point is a stationary point of the original nonconvex problem. Users supply each block's solver through a Subproblem class, supporting the distributed optimization research in [1].

Bar charts compare peak memory and solution wall time of monolithic and distributed formulations as the number of cart-pole trajectory copies increases.
Peak memory stays constant for the distributed formulation as trajectory copies are added, while the monolithic formulation's memory grows. [2]

References

  1. Nicholas C. Orndorff, Christopher Lupp, John T. Hwang. A Distributed Algorithm for Large-Scale Multidisciplinary Design Optimization With Global Constraints. AIAA AVIATION 2026 Forum, 2026. Distribution Statement A: Approved for public release; distribution is unlimited. PA# AFRL-2026-2045
    DOIPDF
  2. Nicholas C. Orndorff, John T. Hwang. A Distributed Method for Solving Large-Scale Multidisciplinary Optimization Problems. AIAA AVIATION FORUM AND ASCEND 2025, 2025.
    DOIPDF

Research connections

Related to ALBCD

Methods

Authors of related work