Modeling an adiabatic quantum computer via an exact map to a gas of particles

We map adiabatic quantum evolution on the classical Hamiltonian dynamics of a 1D gas (Pechukas gas) and simulate the latter numerically. This approach turns out to be both insightful and numerically efficient, as seen from our example of a CNOT gate simulation. For a general class of Hamiltonians we show that the escape probability from the initial state scales no faster than |\dot{\lambda}|^{\gamma}, where |\dot{\lambda}| is the adiabaticity parameter. The scaling exponent for the escape probability is \gamma = 1/2 for all levels, except the edge (bottom and top) ones, where \gamma <~1/3. In principle, our method can solve arbitrarily large adiabatic quantum Hamiltonians.