The Role of Multiplicative Complexity in Compiling Low T-count Oracle Circuits

Abstract: We present a constructive method to create quantum circuits that implement oracles $|x\rangle|y\rangle|0\rangle^k \mapsto |x\rangle|y \oplus f(x)\rangle|0\rangle^k$ for $n$-variable Boolean functions $f$ with low $T$-count. In our method $f$ is given as a 2-regular Boolean logic network over the gate basis ${\land, \oplus, 1}$. Our construction leads to circuits with a $T$-count that is at most four times the number of AND nodes in the network. In addition, we propose a SAT-based method that allows us to trade qubits for $T$ gates, and explore the space/complexity trade-off of quantum circuits. Our constructive method suggests a new upper bound for the number of $T$ gates and ancilla qubits based on the multiplicative complexity $c_\land(f)$ of the oracle function $f$, which is the minimum number of AND gates that is required to realize $f$ over the gate basis ${\land, \oplus, 1}$. There exists a quantum circuit computing $f$ with at most $4 c_\land(f)$ $T$ gates using $k = c_\land(f)$ ancillae. Results known for the multiplicative complexity of Boolean functions can be transferred. We verify our method by comparing it to different state-of-the-art compilers. Finally, we present our synthesis results for Boolean functions used in quantum cryptoanalysis.