Hamiltonian Mechanics — Phase Space & Hamilton's Equations

Hamiltonian mechanics takes the same physics you already know and rewrites it in terms of generalized coordinates $q_i$ and conjugate momenta $p_i$. Instead of one second-order equation for position, you get two linked first-order equations that show how a system moves through phase space.

For a system with coordinates $q_i$, momenta $p_i$, and Hamiltonian $H(q,p,t)$, the equations of motion are

In words: $H$ generates the time evolution of $q_i$ and $p_i$ in phase space. The symbols are the partial derivatives of $H$ with respect to momentum and position; $\dot{q}_i$ and $\dot{p}_i$ are their time rates of change.

Why The Equations Take This Form

The structure is not arbitrary. The Hamiltonian is a function of the variables that define the state, and its main job is to generate the equations of motion. Hamilton's equations split the dynamics into two clean pieces: $\dot{q}_i$ comes from how $H$ changes with momentum, and $\dot{p}_i$ comes from how $H$ changes with position. The minus sign on $\dot{p}_i$ is what makes the pair consistent, so that energy-like quantities are conserved when $H$ has the right symmetries.

To see what the equations move through, picture phase space — the space whose axes are the coordinates and their conjugate momenta. For one degree of freedom a state is a point $(q,p)$, so phase space is two-dimensional; for $N$ degrees of freedom it has $2N$ dimensions. Ordinary space tells you where a particle is, but phase space tells you where it is and the momentum paired with that coordinate, so one point captures one complete instantaneous state. As time passes, that point traces a curve, and Hamilton's equations give the direction of motion along it. In many standard examples $H$ also equals the total energy written in terms of coordinates and momenta, though that identification needs conditions — the safer statement is that $H$ generates time evolution and often matches total energy.

Worked Example: 1D Harmonic Oscillator

Take a mass $m$ on an ideal spring with spring constant $k$. Let the coordinate be $q$ and the conjugate momentum be $p$. For this system,

where the first term is kinetic energy in terms of momentum and the second is the spring potential energy. Apply Hamilton's equations:

Those two first-order equations already describe the motion. To recover the familiar second-order form, differentiate $\dot{q} = p/m$ with respect to time:

then substitute $\dot{p} = -kq$:

the usual simple harmonic motion equation. The lesson: Hamiltonian mechanics does not describe a different oscillator — it describes the same physics in a form that is often easier to generalize.

Try It Yourself

Start from

for a particle moving vertically near Earth's surface. Apply Hamilton's equations to find $\dot{q}$ and $\dot{p}$, then compare with the constant-acceleration model you already know. Check your answer this way: $\dot{q} = p/m$ and $\dot{p} = -mg$, so differentiating gives $\ddot{q} = -g$, exactly the free-fall acceleration. If that matches, you have applied the procedure correctly.

Calculation Pitfalls To Watch

Assuming the Hamiltonian is always total energy. True for many standard systems, not as a universal rule. The safer statement: $H$ generates time evolution, and in many common cases it also equals total energy.

Confusing phase space with ordinary space. Phase space includes momentum coordinates as well as position coordinates, so a phase-space point is not just a location.

Treating $p$ as the same as $mv$ in every coordinate system. In simple Cartesian cases momentum often looks like $p = mv$, but in general coordinates conjugate momentum must be defined from the model, not guessed.

Thinking first-order equations are less complete. Hamilton's equations are first-order, but together they hold the same dynamical information as the familiar second-order equations.

When To Reach For It

Hamiltonian mechanics earns its place when you want a structured view of classical dynamics, especially for systems with many degrees of freedom, conserved quantities, or symmetry, and it is the natural bridge toward statistical mechanics and quantum theory. For simpler problems Newton's laws may be the fastest route; the Hamiltonian view pays off most when the geometry of the system matters as much as the force calculation. A useful next step is to compare this page with simple harmonic motion and watch the same oscillator in both Newtonian and Hamiltonian language.