Swarm intelligence applied in green logistics: A literature review
Introduction
Global warming, environment deterioration, and government regulations arouse the awareness of academic researchers and industrial practitioners for considering the “green” strategies in logistics industry (Murphy, 2000). Green Logistics (GL) concerns not only the provision of green products or services to customers, but also the overall logistics flow of items from cradle to grave, together with reverse logistics. Various green activities and operations have been implemented, such as production scheduling and network construction. In order to improve the performance of GL, individual logistics parties not only need to implement green activities and operations by themselves, but also the cooperation and collaboration among different logistics parties (Zhou et al., 2000). The performance of GL cannot be measured simply in an economic way, but in a sustainable way taking account of environmental and societal factors as well, which are also the objectives of GL (Björklund et al., 2012, Hervani et al., 2005). GL can be understood as the combination of traditional logistics and reverse logistics (RL). Traditional logistics comprises the flow from the raw materials to finished products, while RL is a rather new research field, which involves the concept of recycling used products in order to reduce waste and to increase an industry׳s performance and resulting profits. RL is of great importance, as it not only complements integrated logistics research, but also improves the performance of GL significantly in terms of all the economic, environmental and societal objectives (Lee and Lam, 2012). RL also consists of numerous activities and operations, such as returned products collection, examination, pre-processing, recycling, remanufacturing or disposal (Rogers and Tibben-Lembke, 2001).
Most of the activities and operations in GL can be formulated as combinatorial optimization (CO) models with multiple objectives, constraints and decision variables. Exact methods, such as Linear Programming (LP) and Branch-and-Bound (B&B), are becoming less popular for solving CO problems, as they are either unable to solve complicated CO problems with large numbers of variables or it takes long time to find the solution for CO problems (Laporte, 1992). By contrast, meta-heuristic approaches are becoming increasingly popular as these approaches are approximate approaches, which suggest that they could find satisfactory solutions within an acceptable time instead of finding the optimal solution. Intuitively speaking, meta-heuristic approaches can be classified into two categories: the single-solution based approaches and the population based approaches (Blum and Roli, 2003). The single-solution based approaches, also named the trajectory methods (Consoli and Darby-Dowman, 2007), such as Tabu Search (TS), Simulated Annealing (SA) and various local search methods, in that only one candidate solution exists during the whole search process. However, the population based approach indicates that the search process starts with a population of candidate choices, and the whole population further evolves. The advantages and disadvantages of both the single-solution based approaches and the population based approaches can be found in the literature (Glover and Kochenberger, 2003, Jones et al., 2002). Two important examples of the population based approach are Evolutionary Algorithms (EAs) and Swarm intelligence (SI). The most typical example of EAs is the Genetic Algorithm (GA), which was proposed by Holland in 1975 and simulates the Darwin evolution concept (Holland, 1975).
SI approaches were originally inspired by the collective behavior of natural species, such as ant colony optimization (ACO) from ants, Particle Swarm Optimization (PSO) from birds and the Artificial Bee Colony (ABC) from bees (Bonabeau et al., 1999). SI is a relatively new branch of meta-heuristics comparing with EAs and other single-solution based approaches. SI approaches use approximate and non-deterministic strategies to effectively and efficiently explore and exploit the search space in order to find near-optimal solutions (Blum and Li, 2008, Blum and Merkle, 2008). SI has three fundamental and essential properties, namely decentralization, self-organization and collective behavior, which are necessary and sufficient to acquire SI behaviors. Decentralization means that no central control mechanism exists. The behaviors of individuals are determined by themselves. And the self-organization of individual relies upon four fundamental properties, i.e. positive feedback, negative feedback, fluctuations and multiple interactions (Jeanne, 1986). The interaction between two individuals or environment follows simple rules. The result from interaction would either impel or restrain the behavior of certain individual as positive feedback or negative feedback. The decision of certain individual might be affected by some random factors, which leads to fluctuations. Collective behavior refers that in a swarm, the individual behavior may act randomly, however the aggregation of individual behavior turns to be globally intelligent. In other words, SI indicates that a number of cognitive individuals accumulate their knowledge through the interaction with other individuals or the environment, determine their behaviors solely and finally achieve the target. The characteristics and details of each SI approach are presented in Section 4.
Ever since the introduction of SI, various SI algorithms have been proposed and applied to solve the CO problems in multifarious disciplines, among which the domain of GL could be a promising research area due to its inherent characteristics and features. However, given that there have been many researches of solving GL problems using SI algorithms. Most of them are individually separated, either solving an independent GL problem or adopting a single SI algorithm. In this regard, a comprehensive and extensive literature review of the integration of GL and SI from both the problem context and the methodology perspective is needed urgently. In this research, the state-of-the-art applications of the SI algorithms in the GL background are fully investigated and analyzed, and can help researchers to obtain an intuitive and profound understanding of current research situations. In addition, considering the implementation of various SI algorithms and their variants, this research also provides innovative and universal principles of algorithm selection, improvement and even customization through detailed algorithm analysis and comparison, which can offer practical guidance when solving CO problems using SI related algorithms.
After a brief introduction, the research methodology, i.e. the process of literature review, is described in Section 2. The classification schemes of GL and SI are presented in 3 GL classification scheme, 4 Classification of swarm intelligence respectively. Section 5 discusses the integration of GL and SI and the guiding principles of choosing and optimizing algorithms for specific problems is presented Finally, Section 6 concludes the work and suggests research opportunities and directions for further work.
Section snippets
Research methodology
The objective of this research is to identify major works on interdisciplinary research in GL and SI, and thereafter, to classify and integrate them so as to discover gaps, critical issues and opportunities for further study and research. The literature review is a valid approach and necessary step in exploring new research directions and forms an integral part of the related research, which also helps to scrutinize the conceptual aspects and guides the research toward new theoretical
GL classification scheme
The concept of GL can be well understood from different perspectives. For instance, from the logistics parties point of view, all logistics parties including suppliers, manufacturers, distribution centers and customers are involved in GL; from the standpoint of logistics flows, GL can be viewed as the combination of forward logistics and reverse logistics. The activities and operations in GL, associated with one or more parties, work collaboratively and cooperatively to construct an integrated
Classification of swarm intelligence
The concept of SI was originally introduced by Gerardo Beni and Jing Wang in 1989 in the context of cellular robotic systems (Beni and Wang, 1993), while Bonabeau et al. (1999) redefined it in 1999 as “any attempt to design algorithms or distributed problem-solving devices inspired by the collective behavior of social insect colonies and other animal societies”. Most of the SI approaches are inspired by the collective behavior of natural species, such as ants foraging, birds flocking and bees
Results discussion and implication
In this section, we analyze the results of reviewed publications first with the predefined schemes for both GL and SI. Practical implications and potential research opportunities can be derived from the results. After the results analysis, we provide a mapping scheme between problem types and swarm intelligent algorithms, which can help researchers to get an intuitive and intrinsic understanding of the problems and methods. The useful principles of selection, improvement, hybridization and even
Conclusions and further work
Because of the deterioration of environment and the consumption of the finite and diminishing energy sources, green logistics is gaining increasing attention substantially. Economic performance is no longer the only objective in logistics; two other aspects, environmental and societal performance, are becoming more important than ever before for the purpose of sustainable development. The operational level activities in logistics are the fundamental for any upper management strategies, thus the
Acknowledgments
This work is supported by the Hong Kong Polytechnic University. Our gratitude is also extended to the research committee and the Department of Industrial and Systems Engineering of the Hong Kong Polytechnic University for support in this project (#4-RTY0).
References (194)
- et al.
Optimization of the material flow in a manufacturing plant by use of artificial bee colony algorithm
Expert Syst. Appl.
(2013) - et al.
An ant colony system for enhanced loop-based aisle-network design
Eur. J. Oper. Res.
(2010) - et al.
An ant colony algorithm hybridized with insertion heuristics for the time dependent vehicle routing problem with time windows
Comput. Oper. Res.
(2011) - et al.
Modeling the problem of locating collection areas for urban waste management. An application to the metropolitan area of Barcelona.
Omega
(2006) - et al.
Ant colony optimization techniques for the vehicle routing problem
Adv. Eng. Inf.
(2004) - et al.
The vehicle routing problem in field logistics: Part II
Biosyst. Eng.
(2010) A new saving-based ant algorithm for the vehicle routing problem with simultaneous pickup and delivery
Expert Syst. Appl.
(2010)- et al.
Combining Lagrangian heuristic and Ant Colony System to solve the Single Source Capacitated Facility Location Problem
Transp. Res. Part E: Logist. Transp. Rev.
(2008) - et al.
Activity assigning of fourth party logistics by particle swarm optimization-based preemptive fuzzy integer goal programming
Expert Syst. Appl.
(2010) - et al.
Time dependent vehicle routing problem with a multi ant colony system
Eur. J. Oper. Res.
(2008)
Ant colonies for the traveling salesman problem
BioSystems
A holistic approach for selecting a third-party reverse logistics provider in the presence of vagueness
Comput. Ind. Eng.
The vehicle routing problem: A taxonomic review
Comput. Ind. Eng.
Quantitative models for reverse logistics: a review
Eur. J. Oper. Res.
Ant colony optimization for the two-dimensional loading vehicle routing problem
Comput. Oper. Res.
An ant colony system (ACS) for vehicle routing problem with simultaneous delivery and pickup
Comput. Oper. Res.
Multi-ant colony system (MACS) for a vehicle routing problem with backhauls
Eur. J. Oper. Res.
An ant-colony optimization algorithm for minimizing the completion-time variance of jobs in flowshops
Int. J. Prod. Econ.
Krill herd: a new bio-inspired optimization algorithm
Commun. Nonlinear Sci. Numer. Simul.
A hybrid discrete particle swarm optimization for vehicle routing problem with simultaneous pickup and delivery
Comput. Ind. Eng.
Ant colony optimization for solving an industrial layout problem
Eur. J. Oper. Res.
A modified ant colony optimization algorithm for multi-item inventory routing problems with demand uncertainty
Transp. Res. Part E: Logist. Transp. Rev.
Multi-objective meta-heuristics: an overview of the current state-of-the-art
Eur. J. Oper. Res.
Application of particle swarm intelligence algorithms in supply chain network architecture optimization
Expert Syst. Appl.
Artificial bee colony algorithm for solving sequence-dependent disassembly line balancing problem
Expert Syst. Appl.
Production planning of a hybrid manufacturing–remanufacturing system under uncertainty within a closed-loop supply chain
Int. J. Prod. Econ.
A genetic algorithm-based heuristic for the dynamic integrated forward/reverse logistics network for 3PLs
Comput. Oper. Res.
The vehicle routing problem: An overview of exact and approximate algorithms
Eur. J. Oper. Res.
A bi-objective optimization of supply chain design and distribution operations using non-dominated sorting algorithm: a case study
Expert Syst Appl.
An enhanced ant colony optimization (EACO) applied to capacitated vehicle routing problem
Appl. Intell.
Managing reverse logistics to enhance sustainability of industrial marketing
Ind. Market. Manag.
A discrete artificial bee colony algorithm with composite mutation strategies for permutation flow shop scheduling problem
Sci. Iran.
An ant colony optimization algorithm for setup coordination in a two-stage production system
Appl. Soft Comput.
Development of new features of ant colony optimization for flowshop scheduling
Int. J. Prod. Econ.
A hybrid discrete artificial bee colony algorithm for permutation flowshop scheduling problem
Appl. Soft Comput.
A Hybrid Ant Colony System for Green capacitated vehicle routing problem in sustainbale transport
J. Theor. Appl. Inf. Technol.
A PSO-based optimum consumer incentive policy for WEEE incorporating reliability of components
Int. J. Prod.Res.
A collaborative ant colony algorithm to stochastic mixed-model U-shaped disassembly line balancing and sequencing problem
Int. J. Prod. Res.
Artificial bee colony algorithm for large-scale problems and engineering design optimization
J. Intell. Manuf.
A new heuristic optimization algorithm for modeling of proton exchange membrane fuel cell: bird mating optimizer
Int. J. Energy Res.
Swarm intelligence: application of the ant colony optimization algorithm to logistics-oriented vehicle routing problems
J. Bus. Logist.
Swarm Intelligence in Cellular Robotic Systems, Robots and Biological Systems: Towards a New Bionics?
Performance measurements in the greening of supply chains
Supply Chain Manag.
Swarm intelligence in optimization
Swarm intelligence: introduction and applications
Metaheuristics in combinatorial optimization: Overview and conceptual comparison
ACM Comput. Surveys (CSUR)
Swarm intelligence: from natural to artificial systems
Solving the probabilistic TSP with ant colony optimization
J. Math. Model. Algor.
Organizational research methods: A guide for students and researchers
Cited by (110)
A systematic literature network analysis of green information technology for sustainability: Toward smart and sustainable livelihoods
2024, Technological Forecasting and Social ChangeUtilizing hybrid metaheuristic approach to design an agricultural closed-loop supply chain network
2023, Expert Systems with ApplicationsA hybrid Artificial Immune optimization for high-dimensional feature selection
2023, Knowledge-Based SystemsMGA-IDS: Optimal feature subset selection for anomaly detection framework on in-vehicle networks-CAN bus based on genetic algorithm and intrusion detection approach
2022, Computers and SecurityCitation Excerpt :Meta-heuristic algorithms are used to solve complex optimization challenges and NP-hard problems. These algorithms are suboptimal algorithms, and they find adequate solutions in an acceptable time, instead of finding exact solutions (Zhang et al., 2015). Intrusion detection datasets contain a lot of irrelevant and redundant features.
Adoption of circular economy practices in small and medium-sized enterprises: Evidence from Europe
2022, International Journal of Production Economics