Abstract
Purpose
Facing the diverse needs of large-scale customers, based on available railway service resources and service capabilities, this paper aims to research the design method of railway freight service portfolio, select optimal service solutions and provide customers with comprehensive and customized freight services.
Design/methodology/approach
Based on the characteristics of railway freight services throughout the entire process, the service system is decomposed into independent units of service functions, and a railway freight service combination model is constructed with the goal of minimizing response time, service cost and service time. A model solving algorithm based on adaptive genetic algorithm is proposed.
Findings
Using the computational model, an empirical analysis was conducted on the entire process freight service plan for starch sold from Xi'an to Chengdu as an example. The results showed that the proposed optimization model and algorithm can effectively guide the design of freight plans and provide technical support for real-time response to customers' diversified entire process freight service needs.
Originality/value
With the continuous optimization and upgrading of railway freight source structure, customer demands are becoming increasingly diverse and personalized. Studying and designing a reasonable railway freight service plan throughout the entire process is of great significance for timely response to customer needs, improving service efficiency and reducing design costs.
Keywords
Citation
Wang, Y. (2023), "Research on design methodology for railway freight service combination plans to meet diverse demands", Railway Sciences, Vol. 2 No. 4, pp. 525-538. https://doi.org/10.1108/RS-10-2023-0039
Publisher
:Emerald Publishing Limited
Copyright © 2023, Yao Wang
License
Published in Railway Sciences. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode
1. Introduction
The transition in China's economy, from rapid growth to high-quality development, has resulted in frequent changes in the structure of freight demands. As a consequence, the logistics demands for energy-intensive industries, such as social bulk commodities, have slowed down, while the transport demands for e-commerce, consumer goods and cross-border logistics have experienced rapid growth. Freight demands are gradually shifting from the previous model of “large-volume transport in few varieties and at low frequencies” to “small-volume transport in multiple varieties and high frequencies”. This shift necessitates the provision of quality services that prioritize higher speeds, punctuality, economical efficiency and a wider variety. To meet the logistics demands of enterprise customers and enhance the quality of railway freight services, an imperative step is the organic integration of railway freight services with modern logistics services, which aims to broaden the service scope and offer customers comprehensive and customized freight services. Facing the diverse freight demands of customers, railway freight services need to be upgraded to door-to-door services, which requires changes in the following two aspects. First, the service chain needs to be extended towards both ends of the “door-to-door” process to satisfy customer needs with enriched service options and cover the door-to-door freight service. The second is operation coordination among various service units, including transport, warehousing, packaging, processing, sorting, delivery and other service functions, which should be supported by efficient and integrated freight service combination plans, ensuring balanced and synchronized control over cost, time and quality. With the increasing railway development empowered by digital, networked and intelligent technologies in China, to cater to the vast array of diversified and personalized demands of customers, an Internet-based deployment of numerous traditional and independent railway freight service resources, which leverages the 95306 network freight service platform, enables real-time dynamic, efficient and intelligent combinations of freight service resources through web service composition and allows for the design of service plans ensuring prompt response, thereby paving a new path toward upgrading railway freight services to a door-to-door pattern.
Given the multitude of freight requests and the wealth of available freight service resources on the 95306 platform, new requirements have arisen concerning the design method and efficiency of railway freight service plans intended to address diverse demands. Among the research literature available, Zhang Fei et al. have developed a logistics-oriented web service composition model and introduced an improved NSGA-II algorithm for solving, for which the effectiveness has been demonstrated (Zhang & Zhang, 2018). Tang Li et al. have constructed an objective model incorporating minimal factors such as service time, cost and carbon emissions by introducing the concept of modular design (Tang & Ling, 2019). The proposed method has been verified through a case study. Shi Xiupeng has proposed that railway freight should prioritize the development of an e-commerce freight service pattern, and has designed a service plan in line with this pattern (Shi, 2012). Chen Dengke has raised the concept of a “one-stop” service pattern for railway freight, using the freight services provided by China Railway Shanghai Group Co., Ltd. as an example (Chen, 2015). Nevertheless, the dearth of research concerning the theories and methods pertaining to the formulation of freight service plans has resulted in inadequate support for the intelligent design of such plans.
Drawing from existing research outcomes and taking into account the diverse demands of customers, this paper introduces a model of railway freight service combination plans, aimed at minimizing response times, service costs, service times and other relevant factors, leveraging separate units of service functions derived from a division of the door-to-door railway freight service system. Additionally, a model solution algorithm is proposed, based on an adaptive genetic algorithm. An empirical analysis was conducted on door-to-door railway freight service plans for transportation from a starch factory in Xi'an to Chengdu. The results demonstrate that the proposed optimized model and algorithm are practical for guiding the design of freight plans in a reasonable and effective manner. Furthermore, they are suitable for application as technical support in promptly responding to diverse customer demands for door-to-door freight services.
2. Overview of railway freight service combination
Modern railway freight services incorporate logistics elements by placing product transportation as the central focus and extending the service chain, to facilitate the shift from “station-to-station” transport to “door-to-door” services. Furthermore, the “door-to-door” service chain offers a range of value-added logistics services to deliver door-to-door freight services to customers. The activities of door-to-door railway freight transport services can be categorized into three modules. As shown in Figure 1, the first module focuses on offering fundamental transport services that revolve around railway block trains. The second module involves expanded services such as warehousing, delivery, loading and unloading. The third module encompasses an organic integration with financial, commercial and consulting services to deliver value-added services, including distribution processing, loading reinforcement and commodity trading.
2.1 Service combination process
Oriented to the diversified and personalized demands of customers and relying on railway service resources, numerous elements are effectively combined from a series of sets such as railway service providers, transport capacity resources, service options and service values according to specific rules or methods to generate freight service combination plans that integrate transport services, extended services and value-added services. During the stage of generating a railway freight service combination plan, a search and matching procedure is performed across multiple sets of service resources based on each freight request. This process, as illustrated in Figure 2, leads to the identification of candidate sets of freight services. Feasible service combination plans are then derived from these candidate service sets, utilizing a quality assessment model. Finally, the optimal plan is selected from these feasible plans.
2.2 Service assignments
A freight service assignment corresponds to a specific service combination. The process of addressing the freight demands from each customer request at the 95306 platform is referred to as a freight service assignment.
2.3 Service functions
Each freight service function is intended to meet a category of customer demands, including transport, distribution, loading and unloading and warehousing. The set of service functions is expressed as
2.4 Candidate service resources
The service resources corresponding to each service function are referred to as candidate freight services. For example, the loading and unloading service function module corresponds to a range of loading and unloading resources such as forklifts, cranes, gantry cranes and truck cranes, which are expressed as
2.5 Service combination plan
A function
2.6 Performance evaluation of service combination plan
The performance of freight service combination plans is determined by the attributes of quality evaluation indicators. The generated service combinations are evaluated from the aspects including service cost, service time and service response level (Guo & Wang, 2012; Liu, Song, & Xue, 2013). This evaluation not only relates to the execution quality of individual freight service assignments but also takes into account the combination structure of door-to-door service assignments, which can be incorporated as an objective function of the model.
3. Construction of objective model
The railway freight service combination can be approached as a multi-objective optimization problem.
3.1 Objective model
The set of railway freight service functions is expressed as
3.2 Constraint conditions
Railway freight service combination is composed of service process combination and service function combination. Therefore, the logical structure constraints and quality evaluation constraints of a combination are the focuses of service selection.
3.2.1 Logical structure constraints
A freight service combination plan is composed of multiple sub-service resources by certain rules or methods, and each sub-task corresponds to a group of candidate service sets with the same freight service function but different service values. Generally, the combination plan involves different freight service logic, mainly including sequence, parallelism, selectivity and cyclicity.
Sequence: In the sequential combined freight-service workflow, service functions are executed in sequence, for example, in door-to-door freight services, the pick-up service, loading and unloading service, train service and delivery service are arranged in sequential order.
Selectivity: In the selective combined service, there are multiple selection branches, and the task of each branch is selectively executed with 0-1 decision variables. For example, in trunk transportation, product services from different trains can be selected.
Parallelism: In the parallel combined freight service workflow, there are many parallel branches and the tasks of parallel branches are simultaneously executed, for example, the tasks of financial service module can be executed in parallel with those of the transport module.
Cyclicity: In the cyclic combined service workflow, a task is cyclically executed for k times before proceeding with other tasks. Cycle structure can be regarded as a sequential structure that is executed for multiple times. Therefore, it is not studied herein.
When the
3.2.2 Quality evaluation constraints of combination plan
In practical freight services, customer's expectation arising from their diverse demands for door-to-door freight services becomes the evaluation constraints of service combination plan, which mainly includes treatment time, service time and cost constraints.
Treatment time constraint
The response time
Service time constraint
The total time
Cost constraint
The total expenditure
Selection constraint
Selection identifier constraint of service resources means that only one service resource is selected in the set of selected service function resources, and no service resource can be selected under the service function that is not selected. In this case, the constraint formula is as follows:
The final constraint condition is:
4. Adaptive genetic algorithm
For the multi-objective service combination problem, commonly used particle swarm algorithms, ant colony algorithms, genetic algorithms, etc. (Huang, Wang, & Xie, 2011; Fan, Jiang & Fang, 2010; Īsmail & Koç, 2015; Wang, Liu, & Zhu, 2008; Xia, Cheng, & Chen, 2012), there are often easy to fall into the local optimization and other problems, in this paper, an adaptive genetic algorithm is used to solve the target model.
4.1 Algorithm optimization idea
Existing literature for genetic algorithm improvement has more research (Xu & Chen, 2019; Xiong, Feng, & Chen, 2012; Cha & Yang, 2013; Zhang & Wu, 2012), to meet the extensive and diverse demands of freight transport by railways, with the expansion of search space, challenges are created under a certain population size, such as reduced search efficiency, premature convergence, and local optimization.
The individual selection in a genetic algorithm works by elite optimization strategy, which optimizes two populations at the same time, including the primary population
4.2 Genetic manipulation
Chromosome coding
The service combination plan should select service resources for each service function, so the coding of this plan includes two parts: service function coding and service resource coding. The certificate coding plan of this paper is adopted to develop a bilayer coding system.
In genetic manipulation, each chromosome represents a service combination plan
The first
As shown in Figure 3, the diagonal elements
Chromosome initialization
Determine the freight service functions and control structure that meet diverse demands, and select any freight services that meet local optimal constraints from the candidate resources corresponding to each service function to generate the initial population
Elite strategy
First, the parents are obtained from the current population and the elite population, respectively, and the top k elites are selected from the offspring of each generation into the external population. When the number of individuals in the external population reaches the limit value of 100, the new individuals need to be ranked and selected before entry and those who rank behind 100 will be eliminated and enter into the current population. If the number of current population exceeds 100, ranking and selection should be applied likewise.
4.3 Genetic algorithm steps
In the course of solving railway freight service combination plan model with adaptive genetic algorithm, the following optimized search steps should be used.
5. Case Study
5.1 Case description
A starch company in Shaanxi has five corn starch production lines. It has an annual processing and transformation capacity of one million tons of corn, producing 700,000 tons of corn starch as well as 260,000 tons of by-products such as corn oil, corn protein powder, corn germ meal and corn bran. The company's delivery is made from Huxian Station, which is approximately 6 km away from the company. After the goods arrive at Chengdu North Railway Station, the average distance from Chengdu North Railway Station to each distribution point is about 25 km. The railway distance from Huxian Station to Xi'an West Railway Station is 42 km, and the railway distance from Xi'an West Railway Station to Chengdu North Railway Station is 842 km. For pallet traffic, the pallet can carry 1 ton of goods, while the freight volume for covered goods wagon is 60 tons/wagon and for 40-foot containers is 27 tons/container. The specific freight service requirements of the company are shown in Table 1.
Based on the aforementioned freight demands of corporate customers, the transport, extended and value-added freight service resource set available at the corresponding station sections of China Railway Xi’an Group Co., Ltd. and China Railway Chengdu Group Co., Ltd. are presented in Table 2
Taking into account the available freight service resources at the corresponding station sections of China Railway Xi’an Group Co., Ltd. and China Railway Chengdu Group Co., Ltd., for railway loading and unloading operations, four fork-lift trucks can be deployed to work in pairs, transporting 0.9 tons of goods by relay forklifting, ensuring that the goods do not touch the ground before being loaded into covered goods wagons or containers. For loading operations, two sets of reach stackers can be deployed. For railway transport products, there are options available including container trains and ordinary covered goods wagon trains. Regarding warehousing, CR-Chengdu provides options of flat warehouses and three-dimensional warehouses at the destination station. As for loading reinforcement materials, based on the starch packaging specifications, there are two options available: flexible freight containers and pallets. In addition, there are two options for transporting goods from the starch factory to the station. The first option is to transport goods by container trucks in a short distance to Xi'an West Railway Station, and the second option is to transport goods by container trucks in a short distance to Huxian Station, and then transfer them by railway to Xi'an West Railway Station for consolidation. Relatively optimal solutions for railway freight service plans is shown in Table 3.
5.2 Solution results
By combining the actual data of the case, an adaptive genetic algorithm is applied to solve the built service combination optimization model. The relevant parameter settings are as follows: three objective functions, 87 decision variables, a population size of 20, 100 iterations, a crossover probability of 0.9 and a mutation probability of 0.1.
After conducting the solution computations, three relatively optimal solutions for railway freight service plans are obtained, thereby excluding other feasible combination plans. Plan 1 has the shortest service time and the shortest response time, with a door-to-door service time of only 56.6 h. However, it has the highest cost. Plan 2 has the lowest cost, with a door-to-door service cost of nearly RMB 540,000, but it has the longest door-to-door service time and the longest response time. Plan 3 is a compromise plan, with the service cost, service time and response time falling in the middle range. For the three plans, in addition to the above three indicators, other factors such as customer satisfaction, freight damage and shortage rate, punctuality rate, service resource reliability, service accessibility and resource conservation level are considered and fuzzy comprehensive assessment method is adopted during evaluation and selection of the plans. After the calculation, Plan 3 is identified as the optimal railway freight service combination plan with a door-to-door service time of 94.7 hours and a service price of RMB 558,671 and total response time 34 hours. The specific implementation process of the freight service is shown in Table 4.
6. Conclusion
Relying on the 95306 network freight service platform, China's railway freight transportation sector has been providing freight services to the greatest extent. To cater to the vast array of diversified and personalized demands of customers, new requirements have arisen concerning the design method and efficiency of railway freight service plans. The research on the design methods of railway freight service combination plans aims to promptly respond to customer needs and provide technical support for the design of railway freight service plans. Based on a division of the service system into separate units of service functions depending on characteristics of the door-to-door railway freight service model, this paper presents a model, that utilizes these units to create railway freight service combination plans with minimized response times, service costs, service times and other relevant factors, and proposes a model solution algorithm based on an adaptive genetic algorithm. It can achieve rapid generation of door-to-door freight service plans, which effectively addresses the issue of low efficiency in manual configuration. Through comprehensive evaluation of indicators such as the response time, service cost and service time for the freight service plans, the overall efficiency and benefit of the plan are optimized.
Figures
Figure 1
Railway freight service process oriented to diverse demands
Figure 2
Railway freight service combination process oriented to diverse demands
Figure 3
Chromosome coding
Customer service requirements of the starch company
| Classification of customer requirements | Specific service requirements | Description of requirements |
|---|---|---|
| Customer attributes | Region | Xi'an, Shaanxi Province |
| Delivery destination | Chengdu, Sichuan Province | |
| Single-batch freight volume | 1,800 tons | |
| Goods category | Food | Be kept away from moisture |
| Basic transport requirements | Transport requirements | No specific limitations on train schedules; bagged packaging should be intact and in good condition |
| Extended service | Pick-up service | The enterprise has no private siding, so short-distance transportation by trucks is required to the loading point of railways |
| Handling | No specific requirements for handling method | |
| Delivery service | Delivery in batches according to sales volume | |
| Warehousing service | Due to limited space at the distributor's site, five-day warehousing service for some goods is required | |
| Loading reinforcement | Loading reinforcement materials as well as containerized equipment are required | |
| Value-added service | Sales packaging | Repackaging is required as different distributors need different specifications and quantities of goods |
| Service attributes | Service time | Except for warehousing services, the door-to-door service time is expected to not exceed 144h |
| Service cost | Except for the warehousing service cost, the door-to-door service cost is expected to not exceed RMB 800,000 | |
| Service response time | The average response time for each service function will not exceed 6h |
Source(s): The author’s own work
Information of candidate service resources
| Service classification | Resource description | Service resource options |
|---|---|---|
| Transport services | Optionality of train types | Express train, ordinary train and container train |
| Extended service | Optionality of pick-up plans | Plan 1: short-distance transportation by container trucks to Xi'an West Railway Station Plan 2: short-distance transportation by container trucks to Huxian Station, followed by railway transfer to Xi'an West Railway Station for consolidation |
| Specificity of delivery plans | Delivery in batches | |
| Optionality of handling tools | Fast operation of fork-lift trucks, and normal speed operation of fork-lift trucks | |
| Optionality of warehouses | Flat warehouses and three-dimensional warehouses | |
| Value-added service | Optionality of loading reinforcement materials | Flexible freight containers and pallets |
Source(s): The author’s own work
Relatively optimal solutions for railway freight service plans
| S/N | Departure end | Train transportation | Arrival end | Value-added service | Total | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Truck loading | Truck pick-up | Truck unloading | Loading and handling | Transfer | Type of trains | Unloading and handling | Truck loading | Truck delivery | Truck unloading | Warehousing | Loading reinforcement | |||
| Plan 1 | Mode selection | Fast operation | Normal speed pick-up | Fast operation | Fast operation | Truck transfer | Container train | Fast operation | Fast operation | Normal speed delivery | Fast operation | Three-dimensional warehouse | Pallet | Total |
| Service time | 8.7 | 1 | 8.7 | 12 | 0.5 | 12 | 12 | 8.7 | 1.5 | 8.7 | – | – | 73.8 | |
| Service cost | 15,057 | 36,000 | 15,057 | 59,295 | 67,000 | 2,95,557 | 59,295 | 15,057 | 41,239 | 15,057 | 55,800 | 27,000 | 7,01,414 | |
| Response time | 4 | 0 | 4 | 2 | 0 | 4 | 2 | 4 | 0 | 4 | – | – | 24 | |
| Plan 2 | Mode selection | Normal speed operation | Normal speed pick-up | Normal speed operation | Normal speed operation | Railway transfer | Ordinary train | Normal speed operation | Normal speed operation | Normal speed delivery | Normal speed operation | Three-dimensional warehouse | Pallet | Total |
| Service time | 15.5 | 1 | 15.5 | 19 | 5 | 28.4 | 19 | 15.5 | 1.5 | 15.5 | – | – | 135.9 | |
| Service cost | 12,443 | 36,000 | 12,443 | 14,416 | 41,798 | 2,56,224 | 14,416 | 12,443 | 41,239 | 12,443 | 55,800 | 27,000 | 5,36,665 | |
| Response time | 6 | 0 | 6 | 4 | 8 | 6 | 4 | 6 | 0 | 6 | – | – | 46 | |
| Plan 3 | Mode selection | Fast operation | Normal speed pick-up | Fast operation | Fast operation | Railway transfer | Normal speed train | Fast operation | Fast operation | Normal speed delivery | Fast operation | Three-dimensional warehouse | Pallet | Total |
| Service time | 8.7 | 1 | 8.7 | 12/ | 5 | 28.4 | 12 | 8.7 | 1.5 | 8.7 | – | – | 94.7 | |
| Service cost | 15,057 | 36,000 | 15,057 | 20,191 | 41,798 | 2,56,224 | 20,191 | 15,057 | 41,239 | 15,057 | 55,800 | 27,000 | 5,58,671 | |
| Response time | 4 | 0 | 4 | 2 | 8 | 6 | 2 | 4 | 0 | 4 | – | – | 34 | |
Source(s): The author’s own work
Optimal implementation process of railway freight service plan
| S/N | Service activities | Service resource options used | Service time (H) | Service cost (RMB) | Response time (H) |
|---|---|---|---|---|---|
| 1 | Truck loading at the departure end | Rapid unloading of container trucks | 8.7 | 15,057 | 4 |
| 2 | Truck Pick-up | Short-distance transportation by container trucks to Huxian Station | 1 | 36,000 | 0 |
| 3 | Truck unloading at the departure end | Rapid unloading of container trucks | 8.7 | 15,057 | 4 |
| 4 | Railway loading | Quick relay loading by four fork-lift trucks | 12 | 20,191 | 2 |
| 5 | Railway transfer | Railway transfer to Xi'an West Railway Station for consolidation | 5 | 41,798 | 8 |
| 6 | Train Transportation | Ordinary covered goods wagon train | 28.4 | 2,56224 | 6 |
| 7 | Railway unloading | Quick relay unloading by four fork-lift trucks | 12 | 20,191 | 2 |
| 8 | Truck loading | Rapid unloading of container trucks | 8.7 | 15,057 | 4 |
| 9 | Truck delivery | Short-distance delivery by container trucks | 1.5 | 41,239 | 0 |
| 10 | Truck unloading | Rapid unloading of container trucks | 8.7 | 15,057 | 4 |
| 11 | Warehousing | Three-dimensional warehouse (warehouse-in/out fees, and storage fees) | – | 55,800 | – |
| 12 | Loading reinforcement materials | Pallet | – | 27,000 | – |
| Total | 94.7 | 5,58,671 | 34 | ||
Source(s): The author’s own work
References
Cha, Y., & Yang, J. (2013). Application of improved ant colony algorithm in cloud computing task scheduling using. Computer Engineering and Design, 34(5), 1716–1719, (in Chinese).
Chen, D. (2015). Exploration of a one-stop service model for railway freight logistics. Shanghai Railway Technology, 2, 14-15. (in Chinese).
Fan, X., Jiang, C., & Fang, X. (2010). Dynamic web service selection based on discrete particle swarm, optimization. Journal of Computer Research and Development, 47(1), 147–156, (in Chinese).
Guo, X., & Wang, Y. (2012). Logistics delivery planning considering logistics service level Multi objective model. Journal of Southwest Jiaotong University, 47(5), 874–880.
Huang, B., Wang, T., & Xie, X. (2011). Extended Qos model-driven logistics services selection approach. Journal of Tsinghua University (Science and Technology, 51(1), 19–24, (in Chinese).
Īsmail, E., & Koç, M. A. (2015). Optimization of a passive vibration absorber for a barrel using the genetic algorithm. Expert Systems with Applications, 42(2), 894–905.
Liu, Z., Song, C., & Xue, X. (2013). Dynamic optimization of logistics services based on situational awareness Combination Research Dao. Computer Engineering and Science, 35(9), 51–57, (in Chinese).
Shi, X. (2012). Research on the service mode of railway logistics E-commerce system. Beijing Jiaotong University. pp. 15-20 (in Chinese).
Tang, L., & Ling, Y. (2019).Design method of railway logistics service solution based on NSGA—I. Railway Transport and Economy,41(7), 8-13. (in Chinese).
Wang, J., Liu, Y., & Zhu, J. (2008). Ant colony algorithm with global adaptive optimization. Journal of Chinese Computer Systems, 29(6), 1083–1087, (in Chinese).
Xia, Y., Cheng, B., & Chen, J. (2012). Optimizing services composition based on improved ant colony algoritnm. Chinese Journal of Computers, 35(2), 270–281, (in Chinese).
Xiong, C., Feng, L., & Chen, L. (2012). Research on task scheduling algorithms based on genetic algorithms in cloud computing. Journal of Huazhong University of Science and Technology (Natural Science Edition), 40(S1), 1–4, (in Chinese).
Xu, Y., & Chen, S. (2019). Research on cloud resource allocation optimization algorithm under the packet cluster mapping framework research. Computer Applications and Software, 36(2), 33–38, (in Chinese).
Zhang, H., & Wu, H. (2012). Cloud resource scheduling based on cellular automata genetic algorithm. Computer Engineering, 38(11), 11–13, (in Chinese).
Zhang, F., & Zhang, J. (2018). Research on logistics service composition based on multi—objective 0ptimization. Joumal of Southwest Jiaotong University, 53(6), 1278–185, (in Chinese).