The Complexity of Simulation
Simulating physical systems can be computationally hard, which is why plasma and quantum systems strain even the largest classical computers.
Simulation as computation
To predict how a physical system behaves, we solve its governing equations numerically on a computer. The cost of that solution is a computational complexity problem: how does the work grow as we demand more resolution, more particles, or more accuracy? For many systems the growth is severe.
Why many bodies are hard
Systems with many interacting parts are expensive because interactions couple everything to everything. Naive pairwise interaction among n particles costs O(n^2) per step; fine time and space resolution multiplies this further. Turbulent and multi-scale phenomena force small steps and fine grids, compounding the cost.
Plasma is a hard case
A magnetically confined plasma, such as in the Hyperion breeder design, couples charged-particle motion, electromagnetic fields, and turbulence across a wide range of length and time scales. Kinetic models track a distribution in position and velocity, a high-dimensional space, and turbulence resists coarse approximation. Full first-principles simulation of a reactor-scale plasma remains beyond routine reach, driving reliance on reduced models and large high-performance computing runs.
- Wide scale separation: from electron gyration to device size
- High-dimensional phase space in kinetic descriptions
- Turbulence with no clean small-parameter to exploit
- Strong nonlinear coupling between fields and particles
Quantum systems are worse
Simulating quantum many-body systems on a classical computer is harder still. The state of n quantum particles lives in a space of dimension exponential in n, so storing the state alone can be infeasible. This exponential wall is exactly what quantum simulation hardness and the promise of quantum computers address.
Coping strategies
Because exact simulation is intractable, science leans on reduced models, approximations, statistical sampling (Monte Carlo), and heuristic closures for unresolved scales. Each trades some fidelity for feasibility. Validating these reductions against experiment and against finer simulations is a central and ongoing task, including for fusion device design.