Why Quantum Systems Are Hard to Simulate
The state of a quantum system grows exponentially with its size, making exact classical simulation infeasible beyond small systems.
The exponential state space
A single quantum bit needs two complex numbers to describe. Two need four, three need eight, and n need 2^n. The state of n quantum particles lives in a space whose dimension doubles with each added particle. At a few hundred particles the numbers exceed any storage a classical computer could ever hold.
Entanglement is the culprit
If particles were independent, we could describe each separately and the cost would grow linearly. Entanglement makes the joint state irreducible: it cannot be factored into per-particle pieces. Correlations bind the whole system together, forcing us to track the full exponential state to be exact.
What classical methods can do
- Tensor networks: efficient when entanglement is low, as in some 1D systems
- Quantum Monte Carlo: powerful but blocked by the sign problem for many systems
- Mean-field and density-functional methods: approximate, trading accuracy for feasibility
The sign problem
For many quantum systems, especially fermions and frustrated magnets, Monte Carlo sampling produces terms that cancel with alternating signs, causing the statistical error to blow up exponentially. This sign problem has no general solution and is one reason certain quantum systems resist even approximate classical simulation.
Feynman's insight
Richard Feynman observed in 1981 that simulating quantum physics on classical computers is fundamentally hard, and proposed using controllable quantum systems to simulate other quantum systems. That idea launched quantum computing. A quantum computer represents the exponential state natively, sidestepping the storage wall.
Connection to fusion and materials
Predicting properties of materials, plasmas, and reacting systems from first quantum principles runs straight into this hardness. It shapes what can be computed classically for reactor materials and plasma chemistry, and it is why quantum simulation, captured by the class BQP, is a leading hoped-for application of quantum machines.