Maritime inventory routing problem with multiple time windows

Nurhadi Siswanto; Stefanus Eko Wiratno; Ahmad Rusdiansyah; Ruhul Sarker

doi:10.3934/jimo.2018091

Article Contents

2019, Volume 15, Issue 3: 1185-1211. Doi: 10.3934/jimo.2018091

This issue Previous Article The optimal pricing and ordering policy for temperature sensitive products considering the effects of temperature on demand Next Article Linear bilevel multiobjective optimization problem: Penalty approach

Maritime inventory routing problem with multiple time windows

1.
Department of Industrial Engineering, Institut Teknologi Sepuluh Nopember, Surabaya 60111, Indonesia
2.
School of Engineering and Information Technology, University of New South Wales, Canberra, ACT 2600, Australia

* Corresponding author: Nurhadi Siswanto
* Corresponding author: Nurhadi Siswanto

Received: May 2017

Revised: February 2018

Early access: June 2018

Published: April 2019

The first author is supported by Ministry of Research, Technology and Higher Education, Republic of Indonesia through International Research Collaboration and Scientific Publication Research Grant No. 536/PKS/ITS/2017.

Abstract Full Text(HTML) Figure(8) / Table(8) Related Papers Cited by

Abstract

This paper considers a maritime inventory routing problem with multiple time windows. The typical time windows considered that certain ports permit ships entering and leaving during the daytime only due to various operational limitations. We have developed an exact algorithm to represent this problem. However, due to the excessive computational time required for solving the model, we have proposed a multi-heuristics based genetic algorithm. The multi-heuristics are composed of a set of strategies that correspond to four decision points: ship selection, ship routing, the product type and the quantity of loading and unloading products. The experimental results show that the multi-heuristics can obtain acceptable solutions within a reasonable computational time. Moreover, the flexibility to add or remove the strategies means that the proposed method would not be difficult to implement for other variants of the maritime inventory routing problem.

Keywords:

Mathematics Subject Classification: Primary: 90B06, 90B10, 90B90; Secondary: 90C11, 68R99.

Citation:

Full Text(HTML)

Figure 1. Loading and unloading activities at a port

Download: Full-size image PowerPoint slide

Figure 2. Daily multiple time windows at a port

Download: Full-size image PowerPoint slide

Figure 3. Detailed activities of a ship during its time in a port

Download: Full-size image PowerPoint slide

Figure 4. Several alternatives of a ship arriving and leaving a port when considering time windows

Download: Full-size image PowerPoint slide

Figure 5. An Example of Chromosome

Download: Full-size image PowerPoint slide

Figure 6. Chromosome in one and two steps

Download: Full-size image PowerPoint slide

Figure 7. Changing states during every assignment in a chromosome

Download: Full-size image PowerPoint slide

Figure 8. The values of the fitness functions for test problem 1 with the 15 day planning horizon from each of 40 runs

Download: Full-size image PowerPoint slide

Table 1. An example of strategies for each decision point

No	Decision point	Strategies
1	Ship selection	Based on the least ships current time
2	Routing	Visit two demand ports with the sequence based on the least CD_ik
3	Loading	Compartment [1] for product[1], compartment[2] for product[2] with loading quantities up to the maximum of compartments capacities
4	Unloading	Divide the same quantities for both ports

| Show Table

DownLoad: CSV

Table 2. Data of port and their storages

No	Description	H₁		H₂		H₃
No	Description	S₁₁	S₁₂	S₂₁	S₂₂	S₃₁	S₃₂
1	Maximum capacity (unit)	160	180	55	41	68	51
2	Minimum level (unit)	0	0	0	0	0	0
3	Initial level (unit)	44	28	19	27	46	25
4	Daily supply/demand rate (unit/day)	8	9	6	4	2	5
5	Fixed setup loading time (day)	0.039	0.059	0.074	0.060	0.067	0.049
6	Variable loading time (day/unit)	0.025	0.010	0.003	0.026	0.028	0.014
7	Fixed setup loading cost (＄)	10	8	6	9	8	10
8	Variable loading cost (＄/unit)	0	0	0	0	0	0
9	Quantity penalty cost (＄/day)	10	10	10	10	10	10
10	Fixed setup port time (day)	0		0		0
11	Fixed setup port cost (＄)	0		0		0
12	Daily starting time windows	7.12am		7.12am		7.12am
13	Daily ending time windows	4.48pm		4.48am		4.48pm

| Show Table

DownLoad: CSV

Table 3. Data of ship and their compartments

No	Description	V₁		V₂
No	Description	C₁₁	C₁₂	C₂₁	C₂₂
1	Maximum compartment capacity	68	31	44	50
2	Initial level	0	0	40	4
3	Current product in the compartment	-	-	P₂	P₁

| Show Table

DownLoad: CSV

Table 4. Data of travelling cost and time between ports

| Show Table

DownLoad: CSV

Table 5. The result of exact algorithm solved using Lingo

Test Problem (TP)	Planning Horizon (PH)	Scenario 1 (M_p=3;M_c=2)		Scenario 2 (M_p=3;M_c=2)		Gap (%)
Test Problem (TP)	Planning Horizon (PH)	Optimal Solution	Run Time (in Second)	Optimal Solution	Run Time (in Second)	Gap (%)
1	10	55.8	1,329	-	﹥﹥8.64E + 4^(*)	-
1	15	91.4	21,423	-	﹥﹥8.64E + 4^(*)	-
2	10	66.8	1,012	66.8	582	0
2	15	103, 0	25,451	103.0	74,166	0
3	10	99.0	46	-	﹥﹥8.64E + 4^(*)	-
3	15	216.0	34,827	-	﹥﹥8.64E + 4^(*)	-
4	10	137.0	640	137.0	211	0
4	15	265.0	47,210	265.0	﹥﹥8.64E + 4⁽ⁿ⁾	0
⁽ⁿ⁾the solution did not terminate before the time limit of 8.64E+4 seconds (24 hours) ^(*)a feasible solution was not obtained before the time limit of 8.64E+4 seconds (24 hours)

| Show Table

DownLoad: CSV

Table 6. The sequence of visiting demand ports

Gene₄	Gene₃	The first visiting port	The second visiting port
0	0	CDP[0]
0	1	CDP[0]	CDP[1]
1	0	CDP[1]
1	1	CDP[1]	CDP[0]

| Show Table

DownLoad: CSV

Table 7. An example of one ship's assignment output

(A) Determining loading and unloading quantities
Product #	Quantity in supply port at the time a ship arrive^(*)		Remaining demands^(*)	Compartment Capacity^(*)		Loading Quantity^(*)		Unload for the first visited port^(*)	Unload for the second visited port^(*)
P₁	40.8		28.6	44.0		28.6		0.0	28.6
P₁	50.4		43.8	50.0		43.8		16.0	27.8
^(*) in product units
(B) Routing and schedule of the selected ship
Source Port	Departure		Destination Port	Port's last visit		Arrival		Waiting time to enter⁽ⁿ⁾	Lay time
Source Port	Date	Time	Destination Port	Date	Time	Date	Time	Waiting time to enter⁽ⁿ⁾	Date	Time
H₃	May 11	2.24pm	H₁	May 7	7.12am	May 12	9.36am	0:00	May 12	9.36am
H₁	May 13	3.37pm	H₃	May 11	2.24am	May 14	10.49am	0:00	May 14	10.49am
H₃	May 15	7.12am	H₂	May 11	7.12am	May 16	7.12am	0:00	May 16	7.12am
⁽ⁿ⁾ in (hour:minutes)
(C) Routing and schedule of the selected ship (continue)
Destination Port	Loading Time(n)		Waiting time for supply/space(n)	Ready to Leave		Waiting time to leave(n)			Leaving time
Destination Port	Loading Time(n)		Waiting time for supply/space(n)	Date	Time	Waiting time to leave(n)			Date	Time
H₁	30:01		0:00	May 13	3.37pm	0:00			May 13	3.37pm
H₃	6:33		0:00	May 14	7.12pm	12:00			May 15	7.12am
H₂	22:38		0:00	May 17	5.50am	1:22			May 17	7.12am
⁽ⁿ⁾ in (hour:minutes)
(D) Information of each visiting port's storage
Destination Port	Product#		CD_ik	Remaining Demand(*)		Level of Storages(*) at ship's lay time		Available space in storage(*)	Loading/Unloading quantity(*)
H₁	P₁		27.30	0.0		40.8		-	28.6
H₁	P₂		26.80	0.0		50.4		-	43.8
H₃	P₁		25.00	0.0		21.1		46.9	0.0
H₃	P₂		21.80	16.0		36.7		14.3	16.0
H₂	P₁		20.23	28.6		23.6		31.4	28.6
H₂	P₂		18.05	27.8		7.0		34.0	27.8
^(*) in product units

| Show Table

DownLoad: CSV

Table Appendix A. The results of the multi-heuristics based GA in comparison to the results of the exact algorithm

Test Problem (TP)	Planning Horizon (PH)	Exact Algorithm Solution	Multi-Heuristics based GA (40 runs repetition)
Test Problem (TP)	Planning Horizon (PH)	Exact Algorithm Solution	No. of Individuals in a population	Best Solution (Min)	Gap (%)	Max. Solution	Average	Standart Deviation	No. of infeasible solutions	Average Running Time (in 2nd)
1	10	55.8	20	55.8	0	55.8	55.8	0	0	50.3
			50	55.8	0	55.8	55.8	0	0	105.6
			100	55.8	0	55.8	55.8	0	0	222.7
	15	91.4	20	91.4	0	108.7	94.3	5.4	0	166.0
			50	91.4	0	102.0	91.9	2.0	0	401.2
			100	91.4	0	91.4	91.4	0	0	879.0
	20	$(*)$	20	107.7	-	148.0	126.9	9.3	0	248.5
			50	107.7	-	135.0	122.2	7.1	0	626.0
			100	107.7	-	122.4	116.0	3.7	0	1,312.0
	25	$(*)$	20	146.3	-	196.9	175.4	14.7	5	234.9
			50	140.8	-	195.4	165.3	15.9	2	774.2
			100	143.4	-	188.7	154.4	10.5	0	1,824.1
2	10	66.8	20	66.8	0	76.8	68.4	3.6	0	93.08
			50	66.8	0	68.6	66.9	0.2	0	212.9
			100	66.8	0	66.8	66.8	0	0	497.1
	15	103.0	20	109.3	6.12	140.2	125.6	6.4	0	197.3
			50	103.0	0	130.5	120.2	6.5	0	535.2
			100	105.2	2.14	124.2	116.6	4.4	0	1,207.4
	20	$(*)$20	149.7	-	191.8	170.3	9.8	3	208.3
			50	142.3	-	184.0	166.2	7.5	5	694.3
			100	149.7	-	191.0	163.7	10.4	3	1,520.2
	25	$(*)$	20	177.5	-	231.2	202.8	12.5	12	295.8
			50	171.6	-	207.3	190.9	10.4	6	938.4
			100	173.6	-	208.7	187.0	8.6	4	1,771.2
3	10	99.0	20	99.0	0	99.0	99.0	0	0	43.9
			50	99.0	0	99.0	99.0	0	0	109.0
			100	99.0	0	99.0	99.0	0	0	239.9
	15	216.0	20	216.0	0	241.0	222.4	5.0	0	147.0
			50	216.0	0	241.0	220.1	4.8	0	377.2
			100	216.0	0	221.0	217.4	2.3	0	796.1
	20	$(*)$	20	306.0	-	423.0	343.6	25.1	0	184.0
			50	304.0	-	344.0	316.5	11.35	0	585.2
			100	304.0	-	401.0	312.0	16.0	0	1,222.6
	25	$(*)$	20	401.0	-	508.0	466.8	24.4	1	236.7
			50	346.0	-	522.0	422.5	46.2	2	753.3
			100	346.0	-	483.0	404.8	41.6	0	1,476.9
4	10	137.0	20	137.0	0	147.0	140.0	4.6	0	84.7
			50	137.0	0	137.0	137.0	0	0	180.4
			100	137.0	0	137.0	137.0	0	0	406.6
	15	265.0	20	277.0	4.53	363.0	291.6	14.6	0	196.3
			50	275.0	3.77	294.0	284.6	5.4	0	530.0
			100	265.0	0	287.0	281.4	4.8	0	1,294.4
	20	$(*)$	20	354.0	-	479.0	423.7	24.7	1	217.5
			50	407.0	-	454.0	423.2	15.2	5	676.3
			100	350.0	-	454.0	396.3	29.5	0	1,498.9
	25	$(*)$	20	484.0	-	632.0	543.2	31.9	14	304.6
			50	431.0	-	567.0	517.9	34.3	7	894.2
			100	431.0	-	558.0	505.3	31.8	6	1,818.1
Note: (^*) a feasible solution was not found before the time limit of 8.64E+4 seconds (24 hours)

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	A. Agra, M. Christiansen, et al., A maritime inventory routing problem with stochastic sailing and port times, Computers & Operations Research, 61 (2015), 18-30 doi: 10.1016/j.cor.2015.01.008.
[2]	F. Al-Khayyal and S. J. Hwang, Discrete Optimization-Inventory constrained maritime routing and scheduling for multi-commodity liquid bulk, Part Ⅰ: Applications and model, European Journal of Operational Research, 176 (2007), 106-130. doi: 10.1016/j.ejor.2005.06.047.
[3]	K. Chakhlevitch and P. Cowling, Hyperheuristics: Recent developments, in Adaptive and Multilevel Metaheuristics (eds. C. Cotta, M. Sevaux, K. Sorensen), Springer-Verlag Berlin Heidelberg, (2008), 3-29
[4]	M. Christiansen and B. Nygreen, A method for solving ship routing problems with inventory constraints, Annals of Operations Research, 81 (1998), 357-378. doi: 10.1023/A:1018921527269.
[5]	M. Christiansen and B. Nygreen, Modeling path flows for a combined ship routing and inventory management problem, Annals of Operations Research, 82 (1998), 391-412. doi: 10.1023/A:1018979107222.
[6]	M. Christiansen, Decomposition of a combined inventory and time constrained ship routing problem, Transportation Science, 33 (1999), 3-16. doi: 10.1287/trsc.33.1.3.
[7]	M. Christiansen and K. Fagerholt, Robust ship scheduling with multiple time windows, Naval Research Logistics, 49 (2002), 611-625. doi: 10.1002/nav.10033.
[8]	M. Christiansen, K. Fagerholt, B. Nygreen and D. Ronen, Chapter 4 maritime transportation in handbooks in operations research and management science, (eds, C. Barnhart, G. Laporte), North-Holland, Amsterdam, 14 (2007), 189-284.
[9]	M. Christiansen and K. Fagerholt, Maritime inventory routing problems, in Encyclopedia of Optimization, Second edition (eds C. A. Floudas, P. M. Pardalos), Springer-Verlag, (2009), 1947-1955
[10]	M. Christiansen, K. Fagerholt, T. Flatberg, Ø. Haugen, O. Kloster and E. H. Lund, Maritime inventory routing with multiple products: A case study from the cement industry, European Journal of Operational Research, 208 (2011), 86-94. doi: 10.1016/j.ejor.2010.08.023.
[11]	K. F. Doerner, M. Gronalt, R. F. Hartl, G. Kiechle and M. Reimann, Exact and heuristic algorithms for the vehicle routing problem with multiple interdependent time windows, Computers and Operations Research, 35 (2008), 3034-3048. doi: 10.1016/j.cor.2007.02.012.
[12]	D. Favaretto, E. Moretti and P. Pellegrini, Ant colony system for a VRP with multiple time windows and multiple visits, Journal of Interdisciplinary Mathematics, 10 (2007), 263-284. doi: 10.1080/09720502.2007.10700491.
[13]	K. C. Furman, J. H. Song, G. R. Kocis, M. K. McDonald and P. H. Warrick, Feedstock routing in the exxonmobil downstream sector, Interfaces, 41 (2011), 149-163. doi: 10.1287/inte.1100.0508.
[14]	A. Hemmati, L. M. Hvattum, et al., An iterative two-phase hybrid matheuristic for a multiproduct short sea inventory-routing problem, European Journal of Operational Research, 252 (2016), 775-788 doi: 10.1016/j.ejor.2016.01.060.
[15]	S. J. Hwang, Inventory Constrained Maritime Routing and Scheduling for Multi-Commodity Liquid Bulk, Dissertation, School of Industrial and Systems Engineering, Georgia Institute of Technology, USA, 2005.
[16]	Y. Jiang and I. E. Grossmann, Alternative mixed-integer linear programming models of a maritime inventory routing problem, Computers & Chemical Engineering, 77 (2015), 147-161. doi: 10.1016/j.compchemeng.2015.03.005.
[17]	J. Lee and B. I. Kim, Industrial ship routing problem with split delivery and two types of vessels, Expert Systems with Applications, 42 (2015), 9012-9023. doi: 10.1016/j.eswa.2015.07.059.
[18]	D. J. Papageorgiou, G. L. Nemhauser, J. Sokol, M. S. Cheon and A. B. Keha, MIRPLib - A library of maritime inventory routing problem instances: Survey, core model, and benchmark results, European Journal of Operational Research, 235 (2014), 350-366. doi: 10.1016/j.ejor.2013.12.013.
[19]	D. J. Papageorgiou and A. B. Keha, et al., Two-stage decomposition algorithms for single product maritime inventory routing, INFORMS Journal on Computing, 26 (2014), 825-847 doi: 10.1287/ijoc.2014.0601.
[20]	V. Rodrigues and R. Morabito, et al., Ship routing with pickup and delivery for a maritime oil transportation system: MIP model and heuristics, Systems, 4 (2016), p31.
[21]	B. Santosa and R. Damayanti, et al., Solving multi-product inventory ship routing with a heterogeneous fleet model using a hybrid cross entropy-genetic algorithm: a case study in Indonesia, Production & Manufacturing Research, 4 (2016), 90-113 doi: 10.1080/21693277.2016.1204961.
[22]	M. Savelsbergh and J. H. Song, Inventory routing with continuous moves, Computers and Operations Research, 34 (2007), 1744-1763. doi: 10.1016/j.cor.2005.05.036.
[23]	N. Siswanto, D. Essam and R. Sarker, Solving the ship inventory routing and scheduling problem with undedicated compartments, Computers and Industrial Engineering, 61 (2011), 289-299. doi: 10.1109/ICCIE.2009.5223771.
[24]	N. Siswanto, D. Essam and R. Sarker, Multi-heuristics based genetic algorithm for solving maritime inventory routing problem, IEEE International Conference on Industrial Engineering and Engineering Management, Singapore, (2011), 116-120 doi: 10.1109/IEEM.2011.6117890.
[25]	J. Sokol, C. Zhang, G. Nemhauser, D. Papageorgiou and M. S. Cheon, Robust inventory routing with flexible time window allocation, Working paper, 2015.
[26]	J. H. Song and K. C. Furman, A maritime inventory routing problem: Practical approach, Computers and Operations Research, 40 (2011), 657-665. doi: 10.1016/j.cor.2010.10.031.
[27]	F. Tricoire, M. Romauch, K. F. Doerner and R. F. Hartl, Heuristics for the multi-period orienteering problem with multiple time windows, Computers and Operations Research, 37 (2010), 351-367. doi: 10.1016/j.cor.2009.05.012.