#building-blocks

Holland's schema theorem states

$\operatorname{E}[m(H, t+1)] \ge \frac{m(H, t)\, f(H)}{a_t}\,(1 - p), \qquad p = \frac{\delta(H)}{l - 1}\,p_c + o(H)\,p_m,$

where $m(H, t)$ is the number of strings in schema $H$ at generation $t$, $f(H)$ their average fitness, $a_t$ the average fitness of the population, $l$ the string length, and $p_c$, $p_m$ the crossover and mutation probabilities. For $H$ = ***1*0*1*** ($l = 11$), $m(H, t) = 20$, $f(H) = 6$, $a_t = 4$, $p_c = 0.5$ and $p_m = 0.01$, what lower bound does the theorem give for $\operatorname{E}[m(H, t+1)]$?