General Relativity Problems And Solutions

Working Through General Relativity Problems Without Losing Your Mind

Most people approaching general relativity problems don't realize how different they are from classical mechanics or even special relativity. The math changes fundamentally. You're not just applying formulas anymore. You're building the framework itself as you go. Here's what actually happens when you sit down with a GR problem. You start by identifying the symmetry of the situation. Spherical? You're probably looking at Schwarzschild. Cylindrical? Maybe something cosmological or a straight cosmic string. Plane symmetry points toward pp-waves or some Friedmann model. This step alone saves most students from three hours of pointless tensor computation.

General Relativity Problems And Solutions

I used to grade problem sets for an upper-level course. The difference between students who finished and those who didn't came down to one habit: writing out the ansatz before touching a single Christoffel symbol. I saw too many people launch into calculating R_mu nu off the top of their heads with a generic metric and end up with seventeen pages of algebra that resolved to zero because they'd made a sign error in the first line. The approach that actually works is straightforward, even if the execution is tedious. Pick your coordinates. Write the line element. List the non-zero metric components. Compute the inverse metric. Then calculate the Christoffel symbols systematically, using the symmetry properties of the lower indices to cut your work in half. One thing nobody tells you about computing Christoffel symbols by hand: use the Lagrangian method when the metric is diagonal. Instead of plugging into the standard formula with partial derivatives, write down the geodesic Lagrangian L = g_mu nu x'^mu x'^nu and take the Euler-Lagrange equations. You get the same connection coefficients but with dramatically fewer intermediate expressions. I switched my graduate students to this method and watching time on problem sets dropped from roughly four hours per set to about two.

The real difficulty in GR problems isn't the computation. It's knowing which computation matters. Beginners will happily calculate every component of the Riemann tensor for a five-dimensional metric when three components tell the whole story. Learning to read the problem and stop when you have your answer is a skill that comes from doing them enough times to recognize the patterns. A concrete example. Consider the classic problem of finding the orbit equation for a test particle in Schwarzschild spacetime. Most textbooks derive the effective potential from the conservation laws and call it a day. But here's where people get stuck in practice: the radial equation contains that extra 1/r^3 term from GR that has no Newtonian analog. When you try to solve it perturbatively, you need to treat that term as a small correction to the Keplerian orbit. The standard approach is to substitute u = 1/r, write u = u_0 + epsilon*u_1 where u_0 is the Newtonian solution, and linearize. The result gives you the perihelion precession formula delta phi = 6*pi*GM/(c^2*a*(1-e^2)). I've seen students miss this entire derivation by trying to integrate the full nonlinear equation directly. Don't do that. The perturbation expansion is the whole point of the exercise. Another common trap involves coordinate singularities versus true curvature singularities. A student once spent two full problem sets computing tidal forces at the Schwarzschild radius r = 2M and concluded that something dramatic was happening there. The Kretschmann scalar K = R_mu nu rho sigma * R^mu nu rho sigma = 48*G^2*M^2/(c^4*r^6) tells a different story. At r = 2M it's perfectly finite. At r = 0 it diverges. That's the only real singularity in Schwarzschild. Learning to check invariants early saves you from chasing ghosts in bad coordinates.

Tools That Actually Help

Working through these problems without any computational assistance is possible but inefficient. Mathematica with the xAct package or SymPy's tensors module will compute Christoffel symbols, Ricci tensors, and curvature scalars for metrics up to about 4x4 without breaking a sweat. I recommend doing the first few problems completely by hand so you understand what the software is doing, then switch to symbolic computation for anything beyond three dimensions or non-diagonal metrics. The manual calculation builds intuition. The software lets you explore geometries you'd otherwise never get through. For numerical work there are a few options. Einstein Toolkit is the standard for numerical relativity simulations but the learning curve is steep and you need a cluster. For simpler things like solving the geodesic equations or visualizing orbital precession, a Python script with SciPy's ODE integrator gets you results in minutes.

What Most Problem Sets Are Actually Testing

The problems in a standard GR course cluster into about five categories and each one rewards a slightly different instinct. Index gymnastics with the metric and its inverses tests whether you're comfortable manipulating tensor equations without getting lost in the subscripts. Computing curvature from a given metric tests patience and systematic organization. Deriving field equations from an action tests your understanding of why GR is formulated the way it is. Solving the geodesic equation for a specific spacetime tests your ability to extract physical predictions. And the harder problems that combine several of these elements test whether you've actually internalized the material or just memorized procedures. I found that the single most effective study habit was to redo every worked example in the textbook covering the key texts by Carroll, Hartle, and Schutz without looking at the solution until you reached the final result. You'll spot gaps in your understanding immediately. The moment you realize you can't reproduce a derivation is the moment you actually learn something. A quick note on what doesn't work. Flashcard apps for GR formulas are mostly useless. The formulas aren't the hard part. Understanding when and why to use each one is. Reading solutions passively without computing them yourself is another common waste of time. I've watched students re-read three chapters of Schutz and then struggle to compute the Ricci tensor for the FLRW metric on a test. The skill is in the doing.

If you're starting out, begin with flat spacetime in different coordinates. Work through how the metric transforms under a Lorentz boost, then under a rotation, then under arbitrary curvilinear coordinates. Calculate the Christoffel symbols for polar coordinates in two dimensions. You should get the same results from the transformation law and from direct computation. This exercise alone makes the tensor formalism feel less abstract and more like a consistent language you can actually use. The rest is practice. The problems are hard because the subject is hard. There isn't really a shortcut around that.