Linear System Co-Design via Lifted Polyhedral Query Solving

Abstract

This paper studies co-design queries on Linear Design Problems (LDPs) — monotone design problems whose feasible functionality–resource relations are polyhedra over Euclidean posets. We show that system-level queries on interconnections of linear components can be solved exactly through a lifted polyhedral formulation that preserves block-angular sparsity, connecting monotone co-design semantics with polyhedral multiobjective optimization.

This is the conference version of the journal-length preprint Scalable Co-Design via Linear Design Problems: Compositional Theory and Algorithms.

Yubo Cai 蔡宇博

PhD student at MIT LIDS, IDSS, and CEE

Ph.D. student at MIT LIDS. Research in optimization, system design, and control.