基于偏好的原油移动路径多目标优化

王舒涵; 堵威; 唐漾; 钟伟民

doi:10.16383/j.aas.c240259

基于偏好的原油移动路径多目标优化

doi: 10.16383/j.aas.c240259 cstr: 32138.14.j.aas.c240259

王舒涵^{1, 2, 3,},
堵威^{1, 2, 3,},
唐漾^{1, 2, 3,},
钟伟民^{2, 3,}

1.
华东理工大学工业控制技术全国重点实验室上海 200237
2.
华东理工大学能源化工过程智能制造教育部重点实验室上海 200237
3.
华东理工大学信息科学与工程学院上海 200237

基金项目: 国家杰出青年科学基金 (61925305), 国家自然科学基金 (62173144, 62203173), 中央高校基本科研业务费专项资金 (222202417006), 上海人工智能实验室资助

详细信息

作者简介:
王舒涵：华东理工大学信息科学与工程学院硕士研究生. 主要研究方向为多目标优化, 生成式进化计算. E-mail: shuhanwang0710@163.com

堵威：华东理工大学信息科学与工程学院副教授. 主要研究方向为进化计算, 大规模优化, 鲁棒优化和复杂工业过程决策优化. 本文通信作者. E-mail: duwei0203@ecust.edu.cn

唐漾：华东理工大学信息科学与工程学院教授. 主要研究方向为智能无人系统和工业智能. E-mail: yangtang@ecust.edu.cn

钟伟民：华东理工大学信息科学与工程学院教授. 主要研究方向为工业过程建模与优化控制. E-mail: wmzhong@ecust.edu.cn

计量
- 文章访问数: 164
- HTML全文浏览量: 67
- PDF下载量: 7
- 被引次数: 0
出版历程
- 收稿日期: 2024-05-11
- 录用日期: 2024-07-04
- 网络出版日期: 2024-07-31
- 刊出日期: 2024-12-20

A Preference-based Multi-objective Optimization for Crude Oil Movement Path

WANG Shu-Han^{1, 2, 3
,},
DU Wei^{1, 2, 3
,},
TANG Yang^{1, 2, 3
,},
ZHONG Wei-Min^{2, 3
,}

1.
State Key Laboratory of Industrial Control Technology, East China University of Science and Technology, Shanghai 200237
2.
Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, East China University of Science and Technology, Shanghai 200237
3.
School of Information Science and Engineering, East China University of Science and Technology, Shanghai 200237

Funds: Supported by National Science Fund for Distinguished Young Scholars (61925305), National Natural Science Foundation of China (62173144, 62203173), Fundamental Research Funds for the Central Universities (222202417006), and Shanghai Artifcial Intelligence Laboratory

More Information

Author Bio:
WANG Shu-Han　Master student at the School of Information Science and Engineering, East China University of Science and Technology. His research interest covers multi-objective optimization and generative evolutionary computation

DU Wei　Associate professor at the School of Information Science and Engineering, East China University of Science and Technology. His research interest covers evolutionary computing, large-scale optimization, robust optimization and complex industrial process decision optimization. Corresponding author of this paper

TANG Yang　Professor at the School of Information Science and Engineering, East China University of Science and Technology. His research interest covers intelligent unmanned system and industrial intelligence

ZHONG Wei-Min　Professor at the School of Information Science and Engineering, East China University of Science and Technology. His research interest covers modeling and optimization control of industrial process

摘要

摘要: 原油移动路径规划是原油调度中至关重要的子任务, 直接影响到生产过程中原油供给的稳定性和付油的高效性. 由于此任务需要考虑大规模罐区内复杂的设备条件, 并受到严格的工业生产约束, 同时需要兼顾途径阀门数量与泵机组运力, 导致目前依然倚重调度人员的人工经验来制定路径规划方案, 对传统算法和进化算法的应用提出了挑战. 据此, 本研究基于有向图结构对大规模原油罐区进行细致数学建模, 并提出一种基于偏好的原油移动路径多目标优化(Preference-based multi-objective optimization for crude oil movement path, PB-MOO)算法, 突破了过去高度依赖人工方法的局限性, 为原油移动路径规划提供智能化解决方案. 实验证明该算法能够在满足实际约束的条件下, 找到复杂任务的高质量候选解, 验证了其在此领域的可行性和有效性.
- 大规模原油调度 /
- 路径规划 /
- 多目标优化 /
- 偏好策略
Abstract: The planning of crude oil movement path is a crucial subtask within crude oil scheduling, directly impacting the stability of crude oil supply and the efficiency of oil delivery in the production process. Given the need to consider complex equipment conditions within large-scale tank areas and strict industrial production constraints, while also balancing the number of valves along the route with pump unit capacity, the current reliance on manual experience of schedulers for developing path planning schemes poses a challenge to the application of traditional algorithms and evolutionary algorithms. Therefore, this study undertakes a meticulous mathematical modeling of large-scale crude oil tank areas based on directed graph structures. It proposes a preference-based multi-objective optimization algorithm for crude oil movement path (PB-MOO), overcoming the limitations associated with past heavily manual methods and providing an intelligent solution for crude oil movement path planning. Experimental results demonstrate that the algorithm can identify high-quality candidate solutions for complex tasks while meeting practical constraints. This validates its feasibility and effectiveness in this field.
- Large-scale crude oil scheduling /
- path planning /
- multi-objective optimization /
- preference-based strategy

HTML全文

图 1 原油调度流程示意图

Fig. 1 Schematic diagram of crude oil scheduling process

下载: 全尺寸图片幻灯片

图 2 简化的原储罐区管道运输流程图模型

Fig. 2 Simplified pipeline transportation flowchart model for crude oil tank areas

下载: 全尺寸图片幻灯片

图 3 PB-MOO算法流程图

Fig. 3 PB-MOO algorithm flowchart

下载: 全尺寸图片幻灯片

图 4 从罐T05和罐T13出发的2个子任务各自偏好泵机权重

Fig. 4 Preference pump weights of two subtasksstarting from tank T05 and T13

下载: 全尺寸图片幻灯片

图 5 交叉算子示意图

Fig. 5 Crossover operator diagram

下载: 全尺寸图片幻灯片

图 6 变异算子示意图

Fig. 6 Mutate operator diagram

下载: 全尺寸图片幻灯片

图 7 2个子任务下的原油移动路径规划

Fig. 7 Crude oil movement path planning under two subtasks

下载: 全尺寸图片幻灯片

图 8 4个子任务下的原油移动路径规划

Fig. 8 Crude oil movement path planning under four subtasks

下载: 全尺寸图片幻灯片

图 9 6个子任务下的原油移动路径规划

Fig. 9 Crude oil movement path planning under six subtasks

下载: 全尺寸图片幻灯片

图 10 6个子任务下的消融实验

Fig. 10 Ablation experiments under six subtasks

下载: 全尺寸图片幻灯片

图 11 6个子任务下的IGD值对比测试

Fig. 11 Comparison test of IGD values under six subtasks

下载: 全尺寸图片幻灯片

表 1 参数说明

Table 1 Parameter descriptions

符号	描述
$BH_b$	第$b $个调合头
$d_{k_1,\;k_2}$	节点$v_{k_1}$与节点$v_{k_2}$ 之间的路径长度
$IL_l$	第$l $条进泵线
K	节点总数
L	进泵线个数
J	罐底阀个数
N	原油类型总数
NB	调合头节点个数
$ND_{k,\;r}$	第$k $个个体中第$r $条路径包含的节点个数
NP	泵节点个数
NT	罐节点个数
$Q_{n,\;r}$	原油移动路径$R_{n,\;r}$中泵能提供的最大流量
$H_{n,\;r}$	原油移动路径$R_{n,\;r}$中泵的额定扬程
$\eta_{n,\;r}$	原油移动路径$R_{n,\;r}$中泵的运行效率
$P_{n,\;r}$	原油移动路径$R_{n,\;r}$中泵的功率
$PP_o$	第$o $个泵
$Q_{n,\;r}^{dmand}$	原油移动路径$R_{n,\;r}$中的需求流量
$Q_{n,\;r}^{pump}$	原油移动路径$R_{n,\;r}$中泵的流量
R	路径$R_{n,\;r}$的节点个数
$TK_t$	第$t $个罐
$VL_j^{TK_t}$	罐$TK_t$的第$j $个罐底阀
$p_k$	表示$v_k$是否为泵节点
$x^k_{n,\;r,\;i}$	第$k $个个体中第$n $个子任务的第$r $个节点为$v_i$
$y_{k,\;IL_l,\;PP_o}$	$v_k$为进泵线$IL_l$与泵$PP_o$之间的阀门
$y_{k,\;PP_o,\;BH_b}$	$v_k$为泵$PP_o$与调合头$BH_b$之间的阀门
$y_{j,\;IN\_TK_t}^k$	$v_k$为连接罐$TK_t$的罐底阀$VL_j^{TK_t}$的进罐阀门
$y_{j,\;OUT\_TK_t,\;IL_l}^k$	$v_k$为罐底阀$VL_j^{TK_t}$与进泵线$IL_l$之间的阀门
$z_{k,\;IL_l}$	$v_k$是否处于进泵线$IL_l$中

下载: 导出CSV

表 2 部分储罐节点信息

Table 2 Information of partial storage tank nodes

罐节点	罐底阀节点	罐容量下限 (t)	罐容量上限 (t)
T01	E-T01-1	4 000	16 000
T01	E-T01-2	4 000	16 000
T02	E-T02-1	4 000	35 000
T02	E-T02-2	4 000	35 000
T03	E-T03	12 000	45 000
T04	E-T04	10 000	16 000
T05	E-T05-1-1	12 000	45 000
	E-T05-1-2
	E-T05-2-1
	E-T05-2-2

下载: 导出CSV

表 4 部分管道节点信息

Table 4 Information of partial pipline nodes

管道名	连接节点	双向边	阀门
8-E-104/T06	T06, N6	否	16 001
10-E-022/T06	4toE-T06, N5	否	11 602
10-E-023/T06	4toE-T06, N1	否	11 603
7-E-102/T06	4toE-T06, N4	否	11 605
E-T06-1	T06, 4toE-T06	是	11 607

下载: 导出CSV

表 3 部分泵节点信息

Table 3 Information of partial pump nodes

泵节点	流量 (m³/h)	扬程 (m)	额定功率 (kW)
P01	47	175	8
P02	100	120	12
P03	200	160	32
P04	500	150	75
P07	800	150	120

下载: 导出CSV

表 5 实验任务设置

Table 5 Experimental task settings

子任务数	起始节点	目标节点	需求流量(m³/h)
2	T03	CDU-B	500
2	T11	CDU-A#	1 100
4	T01	CDU-B	300
	T05	CDU-C	500
	T08	CDU-A	1 300
	T20	E-14去CDU-A#	300
6	T01	CDU-D	300
	T05	CDU-B	500
	T08	CDU-A	1 300
	T13	CDU-A#	1 100
	T20	E-14去CDU-A#	300
	T17	E-12去CDU-A#	500

下载: 导出CSV

表 6 每个任务通过不同算法得到HV值的均值与标准差

Table 6 Each task obtains the mean and standard deviation of HV values through different algorithms

	NSGA-III	RVEA	MOEAD-LWS	PB-MOO
2子任务	9.287e + 2 (6.1e + 1)	8.124e + 2 (1.5e + 0)	8.176e + 2 (9.5e + 1)	1.061e + 3 (3.5e + 1)
4子任务	2.772e + 3 (1.9e + 2)	2.021e + 3 (2.4e + 2)	2.732e + 3 (1.3e + 3)	3.632e + 3 (3.8e + 2)
6子任务	4.275e + 3 (9.5e + 2)	3.726e + 3 (1.9e + 3)	—	5.499e + 3 (1.9e + 2)

下载: 导出CSV

表 7 每个任务通过不同算法得到IGD值的均值与标准差

Table 7 Each task obtains the mean and standard deviation of IGD values through different algorithms

	NSGA-III	RVEA	MOEAD-LWS	PB-MOO
2子任务	7.461e + 0 (3.8e + 0)	9.123e + 0 (5.2e + 0)	1.865e + 1 (6.1e + 0)	3.543e + 0 (3.2e + 0)
4子任务	4.448e + 1 (3.7e + 1)	7.841e + 1 (4.6e + 1)	9.247e + 1 (1.1e + 1)	6.765e + 0 (2.1e + 0)
6子任务	5.167e + 1 (2.3e + 1)	1.054e + 2 (1.5e + 1)	—	4.313e + 0 (1.5e + 0)

下载: 导出CSV

表 8 每个任务通过不同算法获得的最终解

Table 8 The final solution of each task obtained by different algorithms

	NSGA-III	RVEA	MOEAD-LWS	PB-MOO
2子任务	(25, 190.4)	(26, 195.6)	(26, 215.7)	(24, 190.1)
4子任务	(60, 378.2)	(57, 424.5)	(52, 432.7)	(52, 364.3)
6子任务	(90, 500.1)	(93, 522.1)	(91, 522.3)	(89, 450.8)

下载: 导出CSV

参考文献(27)

[1]	黄德先, 江永亨, 金以慧. 炼油工业过程控制的研究现状、问题与展望. 自动化学报, 2017, 43(6): 902−916 Huang De-Xian, Jiang Yong-Heng, Jin Yi-Hui. Present research situation, major bottlenecks, and prospect of refinery industry process control. Acta Automatica Sinica, 2017, 43(6): 902−916
[2]	Zhang W T, Du W, Yu G, He R C, Du W L. Large-scale crude oil scheduling: A framework of hybrid optimization based on plan decomposition. In: Proceedings of 2022 IEEE Congress on Evolutionary Computation. Padua, Italy: IEEE, 2022. 1−8
[3]	Zhang W T, Du W, Yu G, He R C, Jin Y C. Knowledge-assisted dual-stage evolutionary optimization of large-scale crude oil scheduling. IEEE Transactions on Emerging Topics in Computational Intelligence, 2024, 8(2): 1567−1581 doi: 10.1109/TETCI.2024.3353590
[4]	Hou Y, Zhang Y X, Wu N Q, Zhu Q H. Constrained multi-objective optimization of short-term crude oil scheduling with dual pipelines and charging tank maintenance requirement. Information Sciences, 2022, 588: 381−404 doi: 10.1016/j.ins.2021.12.067
[5]	Hou Y, Wu N Q, Zhou M C, Li Z W. Pareto-optimization for scheduling of crude oil operations in refinery via genetic algorithm. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2017, 47(3): 517−530 doi: 10.1109/TSMC.2015.2507161
[6]	Yin J, Huang R, Sun H, Cai S Y. Multi-objective optimization for coordinated production and transportation in prefabricated construction with on-site lifting requirements. Computers & Industrial Engineering, DOI: 10.1016/j.cie.2024.110017
[7]	He C L, Zhang Y, Gong D W, Song X F, Sun X Y. A multitask bee colony band selection algorithm with variable-size clustering for hyperspectral images. IEEE Transactions on Evolutionary Computation, 2022, 26(6): 1566−1580 doi: 10.1109/TEVC.2022.3159253
[8]	Fang Y L, Li Z Y, Wang S W, Lu X W. Multi-objective multi-fidelity optimisation for position-constrained human-robot collaborative disassembly planning. International Journal of Production Research, 2024, 62(11): 3872−3889 doi: 10.1080/00207543.2023.2251064
[9]	Wang R Q, Zhang D M, Kang Z J, Zhou R C, Hui G. Study on deep reinforcement learning-based multi-objective path planning algorithm for inter-well connected-channels. Applied Soft Computing, DOI: 10.1016/j.asoc.2023.110761
[10]	Zhu Z X, Wang F X, He S, Sun Y W. Global path planning of mobile robots using a memetic algorithm. International Journal of Systems Science, 2015, 46(11): 1982−1993 doi: 10.1080/00207721.2013.843735
[11]	Suresh K S, Venkatesan R, Venugopal S. Mobile robot path planning using multi-objective genetic algorithm in industrial automation. Soft Computing, 2022, 26(15): 7387−7400 doi: 10.1007/s00500-022-07300-8
[12]	Zhang W T, Du W L, Du W, He R C, Jin Y C. Large-scale continuous-time crude oil scheduling: A variable-length evolutionary optimization approach. IEEE Transactions on Automation Science and Engineering, DOI: 10.1109/TASE.2024.3380744
[13]	Ramteke M, Srinivasan R. Large-scale refinery crude oil scheduling by integrating graph representation and genetic algorithm. Industrial & Engineering Chemistry Research, 2012, 51(14): 5256−5272
[14]	Panda D, Ramteke M. Reactive scheduling of crude oil using structure adapted genetic algorithm under multiple uncertainties. Computers & Chemical Engineering, 2018, 116: 333−351
[15]	Hou Y, Wu N Q, Li Z W. A genetic algorithm approach to short term scheduling of crude oil operations in refinery. IEEJ Transactions on Electrical and Electronic Engineering, 2016, 11(5): 593−603 doi: 10.1002/tee.22277
[16]	Palakonda V, Kang J M. Pre-DEMO: Preference-inspired differential evolution for multi/many-objective optimization. IEEE Transactions on Systems, Man, and Cybernetics-Systems, 2023, 53(12): 7618−7630 doi: 10.1109/TSMC.2023.3298690
[17]	Deb K, Sundar J. Reference point based multi-objective optimization using evolutionary algorithms. In: Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation. Seattle, USA: ACM SIGEVO, 2006. 635−642
[18]	Zheng J H, Lai N, Guo G Q. $ \epsilon $-pareto dominance strategy based on angle preference in MOEA. Pattern Recognition and Artificial Intelligence, 2014, 27(6): 569−575
[19]	Chaudhuri S, Deb K. An interactive evolutionary multi-objective optimization and decision making procedure. Applied Soft Computing, 2010, 10(2): 496−511 doi: 10.1016/j.asoc.2009.08.019
[20]	Wang C, Jiao L C, Zhao J X, Li L L, Liu X, Liu F, et al. Bi-level multiobjective evolutionary learning: A case study on multitask graph neural topology search. IEEE Transactions on Evolutionary Computation, 2024, 28(1): 208−222 doi: 10.1109/TEVC.2023.3255263
[21]	Deb K, Pratap A, Agarwal S, Meyarivan T. A fast and elitist multiobjective geneticalgorithm: NSGA-II. IEEE Transactions on Evolutionary Computation, 2002, 6(2): 182−197 doi: 10.1109/4235.996017
[22]	Yao Y, Peng Z, Xiao B. Parallel hyper-heuristic algorithm for multi-objective route planning in a smart city. IEEE Transactions on Vehicular Technology, 2018, 67(11): 10307−10318 doi: 10.1109/TVT.2018.2868942
[23]	While L, Hingston P, Barone L, Huband S. A faster algorithm for calculating hypervolume. IEEE Transactions on Evolutionary Computation, 2006, 10(1): 29−38 doi: 10.1109/TEVC.2005.851275
[24]	Zitzler E, Thiele L, Laumanns M, Fonseca C M, Fonseca V G. Performance assessment of multiobjective optimizers: An analysis and review. IEEE Transactions on Evolutionary Computation, 2003, 7(2): 117−132 doi: 10.1109/TEVC.2003.810758
[25]	Deb K, Jain H. An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, part I: Solving problems with box constraints. IEEE Transactions on Evolutionary Computation, 2014, 18(4): 577−601 doi: 10.1109/TEVC.2013.2281535
[26]	Wang R, Zhou Z B, Ishibuchi H, Liao T J, Zhang T. Localized weighted sum method for many-objective optimization. IEEE Transactions on Evolutionary Computation, 2018, 22(1): 3−18 doi: 10.1109/TEVC.2016.2611642
[27]	Cheng R, Jin Y C, Olhofer M, Sendhoff B. A reference vector guided evolutionary algorithm for many-objective optimization. IEEE Transactions on Evolutionary Computation, 2016, 20(5): 773−791 doi: 10.1109/TEVC.2016.2519378