Title: A fast quantum algorithm for computing matrix permanent

Abstract: Computing a matrix permanent is known to be a classically intractable problem. This paper presents the first polynomial time quantum algorithm for computing permanents of arbitrary $N$-dimensional square matrices with multiplicative error and additive error protocols. We transformed the well-known classical algorithm, Ryser's formula, into sums of quantum overlap integrals to develop the quantum algorithm. This requires $\mathcal{O}(N)$ and $\mathcal{O}(N^{3})$ evaluations of the real part of the quantum overlap integrals for real and complex matrices, respectively, with polynomial depth quantum circuits of $\mathcal{O}(N)$ qubits. The existence of the polynomial time multiplicative error estimation of a matrix permanent would strongly imply that quantum computers can access the #P-hard problem. Our quantum algorithm attains the milestone of opening a new direction in quantum computing and computational complexity theory. It could resolve crucial computational complexity issues like locating the bounded error quantum polynomial time (BQP) relative to the polynomial hierarchy, which has been a big complexity question. Depending on the size of a matrix norm of interest, we can also obtain an exponentially small additive error estimation and better estimation against Gurvits' classical sampling algorithm. The additive error protocol is practically implementable in noisy intermediate-scale quantum devices. Moreover, our quantum algorithm directly shows the computational complexity connection between the matrix permanent and the real-time partition function of the quantum Ising Hamiltonian.