Speedup algorithm: error bounds

We use the same procedures as yesterday’s but change the functional settings.

For aggregate model without labor, we set

fc(K,K1)=A*K^alpha-(K1-(1-delta)*K)/Q

For aggegate model with labor, given constant labor Lstar, we set

fc(K,K1)=A*K^alpha*Lstar^(1-alpha)-(K1-(1-delta)*K)/Q

for the same fu(x) respectively.