Optimal Routing for Heterogeneous Fixed Fleets of Multicompartment Vehicles

Table 5

Comparison of RGTS with HFFVRP algorithms on T-8.

T-8	BKS of T-8	HCG^a		HRTR^b		Guided TS^c		TSA^d		RGTS		Gap
		Gap	CPU	Gap	CPU	Gap	CPU	Gap	CPU	OV	CPU
13	1517.84	0.014	473	0	358	0	116	0	56	1534.72	26	1.11
14	607.53	1.335	575	0	141	0	92	0	55	630.38	57	3.76
15	1015.29	0.155	335	0	166	0	79	0	59	1025.87	48	1.04
16	1144.94	0.796	350	0	188	0	130	0	94	1146.38	61	0.10
17	1061.96	0.926	2245	0	216	0	153	0	206	1071.82	102	0.93
18	1823.58	2.554	2876	0	366	0	252	0	198	1860.01	111	2.00
19	1117.51	0	5833	0.253	404	0.253	327	0.253	243	1172.40	301	4.91
20	1534.17	1.669	3402	0	447	0	479	0	302	1574.24	220	2.61
Average	1227.85	2.614	2011	1228.21	286	1228.21	203	1228.21	152	1251.98	115	2.06
Average Gap	0.000	0.931	—	0.032	—	0.032	—	0.032	—	2.060	—	—

aHCG: heuristic column generation by Taillard [4] (SucnSparc workstation 50 MHz).
^bHRTR: record-to-record travel algorithm for HFFVRP by Li et al. [5] (Athlon 1 GHz).
^cGuided TS: guided tabu search by Tarantilis et al. [25] (Pentium IV 2.4 GHz).
^dTSA: tabu search algorithm by Brandão [27] (Pentium M 1.4 GHz).