Polycontrolled PROPs for Qudit Circuits: A Uniform Complete Equational Theory For Arbitrary Finite Dimension
Colin Blake
TL;DR
This work delivers a dimension-uniform, finitary equational theory for unitary qudit circuits by introducing a polycontrolled circuit PROP $\textbf{CQC}_d$ with a primitive control constructor. It constructs a finite axiom system $\mathrm{QC}_d$, uniform in dimension, and proves it sound and complete for the standard unitary semantics via a modular reduction to the linear-optical calculus $\textbf{LOPP}$ using a Gray-code embedding. The core idea is to treat control as a primitive, enabling bounded-arity schemata and a clean completeness proof that generalizes known qubit results to arbitrary finite $d$. The completeness result relies on a robust encoding/decoding pipeline (Gray-code based) and a mimicking argument that transfers all $\textbf{LOPP}$ derivations back into $\mathrm{QC}_d$, yielding a practical, circuit-level rewriting framework for native qudit hardware with resource-efficient compilation and verification potential.
Abstract
We present a finite schematic axiomatisation of quantum circuits over d-level systems (qudits), uniform in every finite dimension d >= 2. For each d we define a PROP equipped with a family of control functors, treating control as a primitive categorical constructor. Using a translation between qudit circuits and the LOPP calculus for linear optics based on d-ary Gray codes, we obtain for each d a finite set of local axiom schemata that is sound and complete for unitary d-level circuits: two circuits denote the same unitary if and only if they are inter-derivable using axioms involving at most three wires. The generators are compatible with standard universal qudit gate families, yielding a sound equational basis for circuit rewriting and optimisation-by-rewriting. Conceptually, this extends the qubit circuit completeness results of Clément et al.\ to arbitrary finite dimension, and instantiates the control-as-constructor approach of Delorme and Perdrix in this setting, while keeping the axiom shapes uniform in d.
