Add random seed to wealth dynamics lecture

[!NOTE]
Triaged from the archiving of the legacy RST repo. Originally filed by @jstac on 2019-12-08 as lecture-python#50. QuantEcon/lecture-python was renamed to lecture-python.rst and archived on 2026-07-16; this issue was moved here because it was verified to still apply to the live lecture. Archiving context: QuantEcon/workspace-lectures#17.

Verified still present 2026-07-16. lectures/wealth_dynamics.md contains no seed, default_rng or Generator anywhere in the file, while update_cross_section is @jit(parallel=True) and draws unseeded np.random.randn() inside a prange loop. The Gini-coefficient-vs-μ_r figure named in the report still changes on every rebuild. For contrast, 43 other lectures in this repo do seed their randomness, so this one is an outlier against the repo's own norm rather than a deliberate convention.

Root cause, already diagnosed in the original thread by @Harveyt47: under prange each thread gets its own thread-local RNG slot, so seeding the main thread only fixes the first batch while the remaining threads reseed from entropy. A tractable fix is per-iteration seeding inside the prange body — Numba supports np.random.seed() in nopython mode — which makes the cross-section deterministic without giving up parallelism.

For example, we want the figure showing Gini coefficient vs mu_r in weath_dynamics.rst to be unchanging.

This is quite subtle due to use of Numba and parallelization (which tends to cause problems with seeds). I'm not sure it can be done but it needs to be investigated and a rewrite of the code might be required.

Start in lectures/wealth_dynamics.md and inspect update_cross_section, especially the @jit(parallel=True) function and its np.random.randn() call inside prange. Rebuild the lecture and compare the Gini-coefficient-vs-mu_r figure across rebuilds; done means the figure is unchanged while parallel execution is retained.