What is the optimal way to measure the galaxy power spectrum?

Measurements of the galaxy power spectrum contain a wealth of cosmological information. In Smith & Marian, we generalized the power spectrum methodology of Feldman et al. to take into account the key tenets of galaxy formation: galaxies form and reside exclusively in dark matter haloes; a given dark matter halo may host galaxies of various luminosities; galaxies inherit the large-scale bias of their host halo. In this paradigm, we derived the optimal weighting scheme for maximizing the signal-to-noise (S/N) on a given band power estimate. For a future all-sky flux-limited galaxy redshift survey of depth b(J) > 22, we demonstrate that the optimal weighting scheme does indeed provide improved S/N at the level of similar to 20 per cent when compared to Feldman et al. and similar to 60 per cent relative to Percival et al., for scales of the order of k similar to 0.5 h Mpc(-1). Using a Fisher matrix approach, we show the cosmological information yield is also increased relative to these alternate methods - especially the primordial power spectrum amplitude and dark energy equation of state. Caveats: uncertainties in cluster masses, non-linear halo bias and redshift distortions may reduce information gains.