Performance Comparison of Mode Choice Optimization Algorithm with Simulated Discrete Choice Modeling

Table 9

Log-likelihood values obtained from four algorithm (exp. 1-2) ().

Algorithms & log-likelihood values	Step size ()
	1/32	1/16		1/8	1/4	1/2	1
DFP-
Starting	-	−869.0964727		−869.0964727	−869.0964727	−869.0964727	-
Convergence	-	−705.5861209		−705.6160738	−705.6565252	−705.6940834	-
Iterations num.	-	79		56	42	29	-
BFGS-
Starting	−869.0964727	−869.0964727		−869.0964727	−869.0964727	−869.0964727	−869.0964727
Convergence	−705.5772393	−705.5855181		−705.5888418	−705.5979521	−705.6348579	−705.6311626
Iteration num.	150	78		49	32	22	20
Algorithms & log-likelihood values	Interval of uncertainty
	LL(0), UL	LL(0), UL	LL(0), UL	LL(0), UL	LL(0), UL	LL(0), UL	LL(0), UL	LL(0), UL
DFP-GSM-
Starting	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727
Convergence	−705.7026844	−705.7337937	−705.7334852	−705.775286	−705.7586887	−705.9187437	−705.8917489	−705.9084545
Iteration num.	23	11	11	10	10	8	8	8
BFGS-GSM-
Starting	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727	−869.0964727
Convergence	−705.5879102	−705.6275413	−705.5961032	−705.6161016	−705.6100727	−705.6006594	−705.6241932	−705.6279414
Iteration num.	15	11	11	10	10	10	9	9