Skip to main content

2023 | Buch

Intelligent Production and Manufacturing Optimisation—The Bees Algorithm Approach


Über dieses Buch

This book is the first work dedicated to the Bees Algorithm. Following a gentle introduction to the main ideas underpinning the algorithm, the book presents recent results and developments relating to the algorithm and its application to optimisation problems in production and manufacturing.

With the advent of the Fourth Industrial Revolution, production and manufacturing processes and systems have become more complex. To obtain the best performance from them requires efficient and effective optimisation techniques that do not depend on the availability of process or system models. Such models are usually either not obtainable or mathematically intractable due to the high degrees of nonlinearities and uncertainties in the processes and systems to be represented. The Bees Algorithm is a powerful swarm-based intelligent optimisation metaheuristic inspired by the foraging behaviour of honeybees. The algorithm is conceptually elegant and extremely easy to apply. All it needs to solve an optimisation problem is a means to evaluate the quality of potential solutions.

This book demonstrates the simplicity, effectiveness and versatility of the algorithm and encourages its further adoption by engineers and researchers across the world to realise smart and sustainable manufacturing and production in the age of Industry 4.0 and beyond.




The Bees Algorithm—A Gentle Introduction
The Bees Algorithm is a popular optimisation method taking inspiration from the food foraging behaviour of honey bees. The algorithm performs a kind of exploitative neighbourhood search combined with random explorative search. This chapter describes in detail the Bees Algorithm and its variants. The description of the Bees Algorithm is framed in the general context of parameter optimisation, highlighting the main issues and how the Bees algorithm addresses them. The state-of-the-art of the empirical and theoretical understanding of the Bees Algorithm is discussed and areas of further work are suggested.
Marco Castellani, D. T. Pham

Manufacturing Process Optimisation

Minimising Printed Circuit Board Assembly Time Using the Bees Algorithm with TRIZ-Inspired Operators
With the increasing use of printed circuit boards (PCBs) in the electronics industry, the assembly time per PCB is critical, as it affects the production time and cost. This research investigates the assembly time per PCB using machines with the moving board with time delay characteristic (MBTD), which involves the complex coordination of a component feeding system, a pick and place system and the positioning movement of the PCB. Many years of research work by different researchers using different optimisation algorithms were applied to obtain the shortest time possible for an MBTD case study on a PCB assembly with 50 component locations with improved results in each case study. This research explores how the Bees Algorithm with TRIZ-inspired operators is applied to this case study to reduce the assembly time to 23.42 s to save significant cost and time when compared to other past research work with different optimisation algorithms.
Mei Choo Ang, Kok Weng Ng
The application of the Bees Algorithm in a Digital Twin for Optimising the Wire Electrical Discharge Machining (WEDM) Process Parameters
In digital manufacturing, Digital Twins (DTs) can be used to represent a physical manufacturing process. In this study the application of the Bees Algorithm (BA) for the optimisation of Wire Electrical Discharge Machining (WEDM) process on a DT was considered. To do this a virtual copy of the product being machined was created to provide an accurate description of the parameters that needed to be modified to complete its production. The proposed approach looks at the measured inputs and outputs from the WEDM process and uses the BA to optimise and achieve the best combination of WEDM process parameters.
Michael S Packianather, Theocharis Alexopoulos, Sebastian Squire
A Case Study with the BEE-Miner Algorithm: Defects on the Production Line
Classification which is used to predict the classes of objects has the disadvantage of ignoring the costs incurred in false predictions. However, wrong predictions can cause different degrees of costs. Therefore, cost-sensitive classification algorithms are in demand in order to improve quality of the classification. In this study, rule-based cost sensitive BEE-miner algorithm which was developed by making use of Bees Algorithm and MEPAR-miner algorithms are used to classify the defects in the production line of a textile company in a considerably better way. When results on the quality defect dataset are analysed, it is observed that BEE-miner algorithm outperforms the MEPAR-miner algorithm in terms of classification cost and accuracy.
Merhad Ay, Adil Baykasoglu, Lale Ozbakir, Sinem Kulluk
An Application of the Bees Algorithm to Pulsating Hydroforming
Pulsating hydroforming is a sheet forming process proposed in the last decade. The numerical simulation of this process requires biaxial stress–strain curves which can be obtained by performing a pulsating hydraulic bulge test. In this study, the input parameters of a pulsating hydraulic bulge test with titanium alloy sheets (Ti-6Al-4 V) are optimised using the Bees Algorithm (BA). The input parameters are amplitude and base pressure; bulge height (h) and minimum thickness (t) at dome apex are outputs. The mathematical modelling of h and the design of an objective function (J) are needed for optimisation. A second-degree polynomial equation is derived for h using curve fitting for three frequencies. Additionally, t is calculated depending on h. The objective function is designed for maximum normalised bulge height and minimum normalised thickness. The results show less thinning at the dome apex with a bulge height similar to that of the traditional monotonous method. Thus, a uniform thickness distribution, which is a critical quality indicator in hydroforming, is obtained with acceptable loss in bulge height. After optimisation, Δt (t-t0) is improved by approximately 9%. The bulge height increases by 15 and 13% in the best experimental case and the BA-optimised results, respectively. Consequently, the ductility of Ti-6Al-4 V sheet is increased, and the input parameters are optimised.
Osman Öztürk, Muhammed Arif Şen, Mete Kalyoncu, Hüseyin Selçuk Halkacı

Production Equipment Optimisation

Shape Recognition for Industrial Robot Manipulation with the Bees Algorithm
Fitting primitive shapes to point cloud scenes is a challenging but necessary step for many robotic manipulation operations. State-of-the-art primitive fitting methods rely on geometric shape estimation or iterative procedures. They are often computationally complex and sensitive to algorithm parameterisation. This study tackles primitive fitting as a parameter optimisation problem, solving it using the Bees Algorithm. The performance of the Bees Algorithm is evaluated on three sets of artificial scenes of varying degrees of blurriness and benchmarked against an evolutionary algorithm. Experimental results proved the precision and consistency of the Bees Algorithm. Primitive fitting times were compatible with real-time application.
Marco Castellani, Luca Baronti, Senjing Zheng, Feiying Lan
Bees Algorithm Models for the Identification and Measurement of Tool Wear
Bio-inspired computing algorithms are emerging approaches that are based on the principles and vision of the biological evolution of nature to implement new and robust competing techniques. Recently, bio-inspired algorithms have been identified in machine learning to find the optimal solutions of problems in production processes. In this framework, swarm intelligence, which is a subfield of artificial intelligence concerning the intelligent actions of biological swarms by the relationship of individuals in such environments, is used to solve problems in the world by simulating such biological behaviours. Swarm intelligence is defined as the development of intelligent algorithms that mimick the behaviour of different animal societies. In particular, the Bees Algorithm displays the foraging behaviour of honeybees to solve optimisation and search problems. The algorithm performs a sort-of exploitative neighbourhood search combined with random explorative search. This chapter describes the use of the Bees Algorithm in its basic formulation for tool wear identification and measurement during turning operations.
Doriana M. D’Addona
Global Optimisation for Point Cloud Registration with the Bees Algorithm
The problem of 3D registration entails the estimation of spatial transformation which best aligns two point sets. Iterative Closest Point is arguably the most popular and one of the most effective algorithms for 3D registration at present. This algorithm uses singular value decomposition to obtain a least squares alignment of two point sets. As a greedy alignment procedure, Iterative Closest Point is liable to converge to sub-optimal solutions. In this study, the problem of 3D registration is addressed using the popular Bees Algorithm metaheuristics. Thanks to its global search approach, the Bees Algorithm is known to be highly impervious to sub-optimal convergence. To increase the efficiency of the search, singular value decomposition is used to exploit the search results of the Bees Algorithm. Experimental evidence showed that the proposed algorithm outperformed Iterative Closest Point in terms of consistency and precision and showed high robustness to noise in the point sets.
Feiying Lan, Marco Castellani, Yongjing Wang, Senjing Zheng
Automatic PID Tuning Toolkit Using the Multi-Objective Bees Algorithm
In this study, studies were conducted within the scope of the tuning of PID control systems, which is very popular in the manufacturing and robotics fields. While optimising the controller parameters, a multi-objective Bees Algorithm (MOBA) method was used to minimise the settling time, rise time, overshoot, and system error all at once. Simulations have been made on an example for the control of robotic systems used in the manufacturing area. As a result of simulations, low settling and rise times with MOBA were achieved, while overshoot was also not allowed. At the same time, the Ziegler-Nichols method and MATLAB PID Toolkit were used to compare the parameters.
Murat Şahin, Semih Çakıroğlu
The Effect of Harmony Memory Integration into the Bees Algorithm
The advent of optimization algorithms facilitated finding good solutions for engineering problems. This paper presents a comparative case study between two algorithms relevant to bee search methods. One of the algorithms was modified by adding harmonic memory, which is a stage of the Harmonic Search Algorithm. Both algorithms were applied to a spherical four-link mechanism for gripper design as a case study. The results in terms of the coupler trajectory of the mechanism showed the superiority of integrating harmony memory into the Bees Algorithm. A prototype is manufactured to show the success of rapid design and production.
Osman Acar, Hacı Sağlam, Ziya Şaka
Memory-Based Bees Algorithm with Lévy Flights for Multilevel Image Thresholding
The Memory-based Bees Algorithm (MBA) is a new optimisation algorithm based on the Bees Algorithm (BA). MBA includes private and social information of honey bees to copy the decision-making capability of the bees. Lévy flights are random processes that are based on a stable distribution called the Lévy distribution. The enhanced so-called Levy MBA (LMBA) is used to reduce the tunable parameters of the basic BA and MBA algorithms. It is tested for Otsu’s multilevel image thresholding method with the peak signal-to-noise ratio (PSNR) as the thresholding quality measurement. The objective is to find optimal threshold values, particularly with the highest quality. The results demonstrated several benchmark problems with the efficiency and robustness of the new algorithm.
Nahla Shatnawi, Shahnorbanun Sahran, Mohamad Faidzul Nasrudin
Α New Method to Generate the Initial Population of the Bees Algorithm for Robot Path Planning in a Static Environment
This research work presents a modified form of the Bees Algorithm for mobile robot path planning. This modification is based on an alternative method to generate the initial population of the Bees Algorithm. The proposed method is adopted with the Bees Algorithm to find the shortest collision-free path for a mobile robot in static environments. The environment is represented using a 2D configuration space method that includes the robot and stationary obstacles—this representation guarantees dealing with a continuous map as a reality. The new approach of initialising the population of the Bees Algorithm ensures finding the initial paths even though the complexity of the given environment. The local search and global search are also implemented to enhance the initial solutions. Several benchmark maps were simulated to compute the fitness of the generated paths. The results obtained using the Bees Algorithm for path planning were compared with those of other algorithms. The simulation proved the significant performance of the Bees Algorithm. The comparison results show the efficiency and superiority of the proposed method in finding the shortest path.
Mariam Kashkash, Ahmed Haj Darwish, Abdulkader Joukhadar

Production Plan Optimisation

Method for the Production Planning and Scheduling of a Flexible Manufacturing Plant Based on the Bees Algorithm
Production planning and scheduling of flexible manufacturing plants are still highly manual labor-intensive tasks. The production efficiency is constrained due to the large number of combinations of feasible machine selection and operation sequence arrangement. In this study, a mathematic model approximating the real working environment and two different Bees Algorithms were compared. In the improved Bees Algorithm with site abandonment technology, different strategies were used for the abandonment of initial sites and elite sites. The simulation results based on actual factory data from Trumpf (China) show that the mathematical model and the Bees Algorithm could help to improve production effectiveness. Moreover, the improved Bees Algorithm with site abandonment technology shows its excellent ability to solve problems such as production planning issues in flexible manufacturing plants.
Chao Wang, Tianxiang Chen, Zhenghao Li
Application of the Dual-population Bees Algorithm in a Parallel Machine Scheduling Problem with a Time Window
The parallel machine scheduling problem with time windows (PMSP-TW) belongs to a category of production scheduling problems. Due to the time window characteristics of each machine, it becomes difficult to solve this problem. For the PMSP-TW problem, we proposed a dual-population Bees Algorithm. Two types of search populations and supplementary populations are set in the Bees Algorithm. The propose of dual populations is mainly aimed at the situation of poor convergence performance of the traditional Bees Algorithm. If the optimization process is inefficient, the population to which the food source belongs will be updated through the population performance evaluation. Experiments can verify the scheduling effect of our proposed algorithm.
Yanjie Song, Lining Xing, Yingwu Chen
A Parallel Multi-indicator-Assisted Dynamic Bees Algorithm for Cloud-Edge Collaborative Manufacturing Task Scheduling
Industrial Internet-of-Things brings cloud and edge resources together to support customized manufacturing. With cloud-edge collaboration, large-scale computational tasks of product and process simulation, force and torque analysis, real-time process control, and so forth, are to be executed in cloud or edge resources, while related manufacturing tasks are to be executed in distributed end devices simultaneously. In this circumstance, hybrid task scheduling becomes a key to implement efficient and intelligent manufacturing. In this paper, a multi-indicator-assisted dynamic Bees Algorithm (MIDBA) is presented to solve large-scale task scheduling problem for cloud-edge collaborative manufacturing. The operators of the Bees Algorithm are modified according to multiple indicators to find suitable cloud-edge collaborative modes, cloud and edge resources. A parallel search scheme is also designed to accelerate the scheduling process for large-scale tasks. We implement numerical studies to examine the proposed algorithm on this problem. Compared to the state-of-the-art algorithms, the parallel MIDBA can find better solutions with lesser time.
Yulin Li, Cheng Peng, Yuanjun Laili, Lin Zhang

Logistics and Supply Chain Optimisation

Bees Traplining Metaphors for the Vehicle Routing Problem Using a Decomposition Approach
In this study, the bees traplining metaphor was adopted for the Bees Algorithm (BA) and the Combinatorial Bees Algorithm (BAC) and applied to solve the vehicle routing problem. The two-parameter Continuous and Combinatorial Bees Algorithms (BA2 and BAC2), equipped with a traplining metaphor intensifier, Bees Routing Optimiser (BRO), were used to solve the capacitated vehicle routing problem with a decomposition approach. In the first phase of the proposed method, the two-parameter Bees Algorithm (BA2) was employed to solve the capacitated facility location problem, resulting in clusters of customers that did not violate the vehicles’ capacity. Then, BAC2 combined with BRO was used to produce the routing plan for each cluster. BA2 and BAC2 implement the traplining foraging technique of bees, which integrates their exploratory and exploitative search mechanisms, to simplify parameter setting and use their threat avoidance tactics to intensify the solution. The results of comparisons with other BA versions indicate that the proposed algorithm improves the accuracy of the basic version by at least 4% while speeding it up fourfold.
A. H. Ismail, D. T. Pham
Supply Chain Design and Multi-objective Optimisation with the Bees Algorithm
Supply chain network design is a complex multi-objective optimisation problem consisting of identifying the best combination of suppliers, manufacturing and transport options inter alia, with the aim of optimising the overall performance of the network. In this chapter, the Bees Algorithm is presented as a powerful tool for designing optimal supply chains by minimising the total cost and the total lead time simultaneously when the number of possible configurations is high, which is classified as a NP-hard problem. The Bees Algorithm shows better performance compared to other well-known approaches and it is effective in solving multi-objective supply chain optimisation problems.
Ernesto Mastrocinque


Collaborative Optimisation of Robotic Disassembly Planning Problems using the Bees Algorithm
Remanufacturing is an important way to realize environmental protection and resource reutilization by reusing end-of-life products. Disassembly is an important process of remanufacturing. Manual disassembly is a common method used to disassemble end-of-life products; it is expensive and inefficient. Recently, robotic disassembly has been proposed to address the shortcomings of manual disassembly. Before the robotic disassembly execution process, robotic disassembly planning, which contains robotic disassembly sequence planning and robotic disassembly line balancing, could be utilized to improve disassembly efficiency. However, the existing research independently studies robotic disassembly sequence planning and robotic disassembly line balancing and ignores their interrelationships. In this chapter, the Bees algorithm is utilized to collaboratively optimize the robotic disassembly planning methods. First, every single robotic disassembly workstation executes the disassembly tasks assigned by using the space interference matrix and robotic workstation assignment method in the robotic disassembly line. Then, the optimization objectives of the collaborative optimization problem are described, and the analytic network process is utilized to assign weights to different objectives. After that, the improved discrete Bees algorithm is used to find the optimal solution, and case studies are carried out to verify the proposed method. In addition, simulations based on RoboDK under different scenarios are also conducted to show the effectiveness of the proposed method.
Jiayi Liu, Quan Liu, Zude Zhou, Duc Truong Pham, Wenjun Xu, Yilin Fang
Optimisation of Robotic Disassembly Sequence Plans for Sustainability Using the Multi-objective Bees Algorithm
In recent years, remanufacturing has become critical for environmental protection and natural resource conservation. The purpose of the work reported in this chapter is to find the best plan for product disassembly, the first step in the recovery of end-of-life products, balancing the three goals of sustainability—economic, energy and environmental. The study proposes three strategies: reuse, remanufacturing and recycling. The Multi-objective Bees Algorithm (MOBA), Non-dominated Sorting Genetic Algorithm II (NSGA II) and Pareto Envelope-based Selection Algorithm II (PESA II) are used to create solutions for two case studies. In this work, MOBA outperforms other algorithms in finding Pareto optimal solutions for robotic disassembly sequence planning in all cases.
Natalia Hartono, F. Javier Ramírez, D. T. Pham
Task Optimisation for a Modern Cloud Remanufacturing System Using the Bees Algorithm
Remanufacturing represents one of the most promising strategies for reaching economic and environmental sustainability goals. The implementation of cloud technologies in the classical remanufacturing process provides an opportunity to define a novel approach called cloud remanufacturing. Referring to this context, this chapter proposes the application of the Bees Algorithm (BA) for task assignation optimisation in the cloud remanufacturing context. Furthermore, a full factorial plan of experiments is proposed to evaluate the influence of different BA parameters on the solutions.
Mario Caterino, Marcello Fera, Roberto Macchiaroli, D. T. Pham
Prediction of the Remaining Useful Life of Engines for Remanufacturing Using a Semi-supervised Deep Learning Model Trained by the Bees Algorithm
The estimation of the remaining useful life (RUL) of a component is one of the most important tasks for predictive maintenance systems (PMSs). This research aims to predict the remaining useful life (RUL) of aircraft engines by using the Bees Algorithm (BA)-optimised semi-supervised deep learning model. To this end, the LSTM architecture has been implemented as a successful deep learning architecture to process time-series datasets. The predictions have been made by using the NASA Turbofan Engine Corruption Simulation (C-MAPSS) dataset. The proposed prediction model has been formulated as a binary classification task based on the semi-supervised labelling process. Unlike the conventional predictive maintenance models, the data-driven LSTM architecture automatically learns features of the multivariate time series. The model has been verified based on sixteen different time cycles. The proposed semi-supervised deep learning model has been trained using the BA to improve the accuracy value for the safe and unsafe conditions of aircraft engines. The experimental results reached 98% accuracy for the test dataset, and the proposed model performed better in terms of the F1 measure for both the training and test datasets. Experiments proved that the proposed deep learning model can be used as a promising RUL prediction model.
Sultan Zeybek
Intelligent Production and Manufacturing Optimisation—The Bees Algorithm Approach
herausgegeben von
Duc Truong Pham
Natalia Hartono
Electronic ISBN
Print ISBN