Text | Mountain people
Edit | Mountain people
«——[·Preface·] ——»
At present, the distribution of agricultural equipment factories in China shows the characteristics of decentralization and miniaturization, resulting in unbalanced utilization of equipment in agricultural equipment manufacturing. Many small agricultural machinery factories still stay in the workshop production stage, with low production efficiency and chaotic management.
In order to improve production efficiency, small agricultural machinery factories have begun to organize production through cloud service platforms. However, a large number of non-standard components that need to be produced on site are used in the manufacturing process of agricultural equipment, and their quality is difficult to quantify, resulting in the difficulty of improving the accuracy and efficiency of agricultural equipment cloud services.
Cloud manufacturing is the product of a combination of advanced information technology, manufacturing technology and emerging Internet of Things technologies. In cloud manufacturing, the processing information of the factory is virtualized and digitized, and encapsulated as a cloud service pool, sharing resources for individual users.
In the cloud manufacturing process, the manufacturing task is decomposed into several subtasks, and each subtask is matched with the corresponding cloud service combination; Then select a specific service. There are obvious mutual constraints between subtasks, so the optimization of subtasks becomes a multi-dimensional problem, which limits the accuracy and efficiency of cloud services.
At present, artificial intelligence algorithms are widely used in complex optimization problems due to their fast optimization speed, such as genetic algorithms (GA), particle swarm algorithms (PSO), MAX-MIN ant systems (MMAS), artificial bee colony algorithms (ABC), firefly algorithms, etc.
Kumar and Bawa proposed a generalized ant colony optimizer (GACO), but GACO has not been tested in specific models, so the utility of GACO needs to be further explored.
With the development of China's agriculture, more and more equipment is used in agricultural production, so it is necessary to improve the accuracy and efficiency of cloud service optimization in agricultural equipment manufacturing. Therefore, a new method of cloud service optimization based on dynamic artificial anti-bee colony algorithm is proposed.
Firstly, the dynamic coefficient strategy and reliability feedback update strategy are applied to the quality of service (QoS) evaluation model to enhance the applicability of the mathematical model in the busy agricultural environment.
Secondly, DAABA is designed based on artificial ant colony algorithm and bee colony algorithm. The traditional artificial bee colony algorithm is added to the mutation operation to improve the accuracy of the optimal solution. The optimal fusion evaluation strategy and iterative adjustment threshold strategy are adopted to improve the accuracy and efficiency of agricultural equipment manufacturing cloud services.
Finally, the convergence of the DAABA algorithm is verified by theoretical derivation and simulation. Experimental simulation verifies that the accuracy and efficiency of DAABA are higher than that of GA, ABC and MMAS.
Mathematical model
1. Cloud service optimization process in agricultural equipment manufacturing
Due to the strong seasonal characteristics of agricultural production, the manufacturing tasks of agricultural equipment are often concentrated in busy farming. In the case of large-scale orders, a reasonable solution can significantly improve the efficiency of the plant.
Therefore, it is necessary to analyze the cloud manufacturing process. As shown in the following figure, the manufacturing task is decomposed into multiple subtasks, and the corresponding cloud service combination is matched for each subtask; The optimal factory is then selected from the corresponding cloud service in a certain order to determine the optimal manufacturing scheme.
The specific process can be divided into the following steps:
(1). Manufacturing task decomposition: Decompose agricultural equipment manufacturing task T into n sub-tasks ST through the cloud platform.
(2). Subtask classification: According to the requirements of processing technology, match the corresponding cloud service combination for each subtask, where CSA represents the cloud service combination containing m factory F.
(3). Subtask matching: After analyzing the manufacturing capacity of m factories in the cloud service combination, the QoS evaluation model is used to match each subtask STi with factory Fi,j, where i represents the ith subtask, i=1, 2,..., j represents the jth factory, j=1, 2,..., n.
2. QoS evaluation model analysis
The quality of service (QoS) evaluation model is often used for various optimization problems, so this paper uses the QoS evaluation model for cloud service optimization. Considering the versatility of the calculation method, the QoS evaluation model is established using only reliability (RE), manufacturing level (ML), scale (SE) and distance (DT) as evaluation indicators.
Reliability (RE). Reliability is mainly determined by the quality of production and the credibility of the factory. In agricultural equipment manufacturing, the higher the reputation of the factory, the higher the reliability. During busy times, agricultural equipment is burdened with heavy production tasks and is prone to failure.
Therefore, in order to ensure agricultural production, the failure rate must be reduced, which puts forward higher requirements for the reliability of agricultural equipment. In order to gain the trust of farmers, priority must be given to factories with high reliability. This part of the data is provided by the factory.
Manufacturing level (ML). The level of music manufacturing is mainly determined by the ability of technicians and the precision of the manufacturing workshop. Factories with higher levels of manufacturing produce better agricultural equipment. This part of the data is provided by the factory.
Scale (SE). Factory size can represent production efficiency. As the size of the plant expands, the time spent in the manufacturing process will be shortened, which is an ideal situation for the manufacturing of agricultural equipment where agriculture is busy. This part of the data is provided by the factory.
Distance (DT). Distance can represent the cost during transportation and is mainly used to solve TSP problems in the manufacturing process. This part of the data is converted from the coordinates provided by the factory.
Reliability feedback updates. A large number of components are used in the manufacturing process of agricultural equipment, which are divided into standard and non-standard parts.
However, non-standard components need to be processed on site, and reliability is affected by external factors such as raw materials and personnel. In order to ensure the quality of agricultural equipment, the reliability of non-standard components must be strictly controlled.
Therefore, a feedback update strategy is adopted to update the reliability of each plant in a timely manner after each round of manufacturing.
For standard parts, the reliability is high and does not need to be considered too much; For non-standard parts, reliability varies greatly, so priority must be given to plants with high reliability.
By adjusting the proportion of reliability in the QoS evaluation model, unreliable factories are selected to further ensure the quality of agricultural equipment. If a fixed factor is used to adjust the proportion of reliability, some discontinuities will occur.
Therefore, the results will be highly volatile. Based on the data analysis, a dynamic coefficient method is proposed to solve this problem. The QoS evaluation model can be expressed as the following formula:
The combined evaluation indicators can be regarded as the aggregation of individual evaluation indicators, and the aggregation method is related to the structure of the subtask. In this article, the manufacturing scenario has a sequential structure, so the specific aggregate functions are shown in the following table.
The QoS value can be calculated by the following equation, where c1, c2, c3, c4 are the values given by the expert:
In general, the solution of cloud service optimization problems based on QoS model is equivalent to finding the maximum value of the QF function. The larger the function value, the better the performance.
Basic algorithm
At present, artificial intelligence algorithms have been widely used in optimization problems and have achieved good results. DAABA is an improved algorithm for artificial colonies and improved algorithms for artificial colony algorithms.
- Maximum-Smallest Ant System (MMAS)
The following table shows the parameter initialization, pheromone initialization, heuristic function, state transition rule, fitness function, and pheromone update of MMAS:
The advantages of MMAS are mainly manifested in the following two aspects:
- Algorithms have the potential for parallelization, resulting in strong robustness and efficiency.
- The algorithm has strong global search capabilities in the early stage.
- Artificial Bee Swarm (ABC) algorithm.
In 2005, Caraboga proposed an artificial bee colony algorithm. With its powerful global search capability, the ABC algorithm has been successfully applied to optimization problems. ABC consists of four essential elements: honey source, hired bees, onlooker bees, and scout bees.
The basic operations of ABC include the honey source fitness function, following probability, and detection, as shown in the following table.
The search scope of the ABC algorithm is determined randomly, which enhances the ability of global search. However, low accuracy in the early stage and slow convergence speed in the later stage are obvious disadvantages.
Dynamic artificial anti-bee swarm algorithm design
Considering the advantages and disadvantages of MMAS and ABC, DAABA is proposed in this paper. Compared with improving the ant colony algorithm and the traditional artificial bee colony algorithm, DAABA has improved in three aspects:
(1) Improve the accuracy of early search. First, DAABA calls MMAS to get a better basic solution. When MMAS stagnates in the later stage, the optimal fusion evaluation strategy is adopted to reduce useless calculation.
(2) Enhance global search capabilities. Compared with the traditional artificial bee colony algorithm, the algorithm with mutation operation can enhance the global search ability.
(3) Optimize the search ability in the later stage.
When the feasible solution is close to the optimal solution, the iterative adjustment threshold strategy is used to increase the number of bee colonies. As a result, the optimal solution can be found more quickly.
- Optimal fusion evaluation strategy
Traditional mixing algorithms control the junction point by setting a fixed number of iterations, even if the convergence effect is not ideal. The algorithm has to perform several useless iterations before calling ABC, which makes mixing algorithms slow. To speed things up, an optimal fusion evaluation strategy is used.
When MMAS stalls, ABC is called in time to avoid waste of computing resources. The specific steps of the optimal convergence evaluation strategy are as follows:
(1) Set the minimum and maximum iterations of MMAS to NAmin and NAmax.
(2) When the conditions shown in the figure below are met, terminate MMAS and call ABC:
- Iteratively adjust the threshold policy
In the early days, DAABA used the powerful global search capabilities of MMAS to quickly converge to the range of optimal solutions. In the later stage, DAABA uses ABC's excellent optimization ability to quickly search for the optimal solution.
The search accuracy of ABC is directly related to the number of bee colonies, and the larger the number of bee colonies, the higher the search accuracy. However, as the number of bee colonies increases, the computational burden increases rapidly.
In order to improve the accuracy of DAABA while reducing the computational complexity, this paper adopts an iterative adjustment strategy. A threshold NBj is designed to adjust for the timing of population growth. NBj can be calculated by the following equation, where NBmax represents the maximum number of iterations and φ represents the adjustment factor.
When the number of iterations of ABC is greater than that of NBj, the number of bee colonies is increased to further find the optimal solution.
Simulations and experiments
5.1. Simulation
In order to ensure that agricultural equipment manufacturing can be completed in time, simulation experiments are required to verify the reliability of DAABA. The specific steps of the simulation are as follows:
(1) Agricultural equipment manufacturing tasks are numbered from 1 to 6. Each task is divided into 10 sub-tasks, including die forging, extrusion, rolling, stretching, cutting, fusion welding, waterjet cutting, plasma cutting, milling, shearing and other processing processes.
(2) The factory number in each cloud service combination is 1 ~ 40, and the manufacturing information is processed according to the evaluation index.
(4) The optimization results of cloud services in agricultural equipment manufacturing are shown in the figure below. The results show that after 200 iterations, DAABA tends to converge, which once again proves the convergence of DAABA.
- Through the task allocation of cloud service pools, multiple agricultural equipment manufacturing tasks can be carried out at the same time. The results of multiple manufacturing tasks are shown in the figure below, where the horizontal and vertical axes represent virtual plane coordinates, each point represents a factory, and each color ring represents a complete agricultural equipment cloud manufacturing scheme.
5.2. Comparative Experiments
In order to verify the accuracy of DAABA in cloud service optimization, experiments based on DAABA, GA, MMAS and ABC were carried out.
The control parameters of each algorithm are shown in the following table.
All algorithms were tested for 9 large-scale problems, and the scale of the test is shown in the following table.
Each algorithm is repeated 20 times to reduce the error, and the optimization result is shown in the figure below, where the task size is expressed as T (n,m), m is the number of subtasks, and n is the number of factories in the cloud service portfolio for each subtask.
To verify the speed of DAABA in cloud service optimization, experiments were conducted on DAABA and ABC. The average time spent by each algorithm is shown in the figure below.
In summary, DAABA's performance is better than several others, and it is more suitable for solving large-scale problems in agricultural equipment cloud service optimization.
conclusion
A large number of non-standard components are used in the agricultural equipment manufacturing process, and their quality is difficult to quantify, which limits the efficiency and accuracy of agricultural equipment manufacturing cloud services.
In order to solve this problem, this paper proposes a cloud service optimization method, puts forward a reliability feedback update strategy and a dynamic coefficient strategy, adjusts the proportion of reliability in the evaluation model, and improves the quality of agricultural equipment.
In order to improve the accuracy and efficiency of cloud manufacturing of agricultural equipment, a dynamic artificial anti-bee colony algorithm (DAABA) is proposed.
In DAABA, the optimal convergence evaluation strategy is used to call ABC in time, which reduces unnecessary iterations, shortens computing time, and improves the operation speed of cloud services.
Add mutation operation to ABC to directly update the honey source to maximize optimization ability.
The iterative adjustment threshold strategy is adopted to increase the number of bee colonies, improve the accuracy of cloud services, and reduce the computational complexity in the later stage.
Finally, the convergence of DAABA algorithm is verified by theoretical derivation and simulation, and the performance of DAABA is verified by experiments based on DAABA, GA, MMAS and ABC.
The results show that the accuracy and efficiency of DAABA in cloud service optimization in agricultural equipment manufacturing industry are better than other methods.
Bibliography:
[1] G. Geng and K. Li,“Research and application of digital design and manufacturing technology of agricultural machinery,”Agro Food Industry Hi-Tech, vol. 28, no. 1, pp. 2891–2894, 2017.
[2] G. Ozogul,“Effffects of changing concepts and conditions of agricultural production to agricultural machinery manufacturing sectorin globalizing world,”Journal of Food Agriculture and Environment, vol. 10, no. 3-4, pp. 456–462, 2012.
[3] B. Ben, R. Ihab, and L. Wang, “Cloud manufacturing: current trends and future implementations,”Journal of Manufacturing Science & Engineering, vol. 137, no. 4, Article ID 044001, 2015.
[4] Y. Cao, S. Wang, L. Kang, C. Li, and L Guo,“Study on machining service modes and resource selection strategies in cloud manufacturing,” Ae International Journal of Advanced Manufacturing Technology, vol. 81, no. 1–4, pp. 597–613, 2015.
[5] Y.-L. Luo, L. Zhang, F. Tao et al.,“A modeling and description method of multidimensional information for manufacturing capability in cloud manufacturing system,”Ae International Journal of Advanced Manufacturing Technology, vol. 69, no. 5–8, pp. 961–975, 2013.
[6] M. Cygan, L. Kowalik, and A. Socala,“Improving TSP tours using dynamic programming over tree decompositions,”European Symposium on Algorithms, vol. 15, no. 4, pp. 1–19, 2019.