Given the existing integrated scheduling algorithms, all processes are ordered and scheduled overall, and these algorithms ignore the influence of the vertical and horizontal characteristics of the product process tree on the product scheduling effect. This paper presents an integrated scheduling algorithm for the same equipment process sequencing based on the Root-Subtree horizontal and vertical pre-scheduling to solve the above problem. Firstly, the tree decomposition method is used to extract the root node to split the process tree into several Root-Subtrees, and the Root-Subtree priority is set from large to small through the optimal completion time of vertical and horizontal pre-scheduling. All Root-Subtree processes on the same equipment are sorted into the stack according to the equipment process pre-start time, and the stack-top processes are combined with the schedulable process set to schedule and dispatch the stack. The start processing time of each process is determined according to the dynamic start processing time strategy of the equipment process, to complete the fusion operation of the Root-Subtree processes under the constraints of the vertical process tree and the horizontal equipment. Then, the root node is retrieved to form a substantial scheduling scheme, which realizes scheduling optimization by mining the vertical and horizontal characteristics of the process tree. Verification by examples shows that, compared with the traditional integrated scheduling algorithms that sort the scheduling processes as an overall, the integrated scheduling algorithm in this paper is better. The proposed algorithm enhances the process scheduling compactness, reduces the length of the idle time of the processing equipment, and optimizes the production scheduling target, which is of universal significance to solve the integrated scheduling problem.

The production scheduling problem has always been the core content and critical technology in manufacturing and composition. It is a constraint satisfaction problem characterized by reasonable use of the limited manufacturing resources and production costs under the condition of satisfying all constraints to ensure the optimal production target selected by the enterprise and maximize efficiency. In general, the solution to the production scheduling problem is NP-hard. It is difficult to form an effective and similar theoretical method to study these scheduling problems. The research of this problem mainly focuses on flow-shop scheduling (flow scheduling for large batches of the same product) [

Nowadays, in the customer-oriented personalized and customized production model, the manufacturing methods based on modular design [

To sum up, this paper presents an integrated scheduling algorithm for the same equipment process sequencing based on the Root-Subtree vertical and horizontal pre-scheduling (ISA-SEPS). Firstly, the algorithm takes component modules or part modules as the basic unit and uses the tree decomposition method of the distributed manufacturing thought to take out the root node. It splits the product process tree into several Root-Subtrees (component modules) to reduce the scale of the process tree. These operations are ready to analyze the vertical and horizontal characteristics of the product process tree. Then, all Root-Subtrees are traversed through the literature [

In the production and processing mode of the integrated scheduling, the complex product process tree model is a directed tree diagram composed of the processes, and these processes have a partial order relationship. The node in the process tree represents the processing or assembly process. There is a one-to-one correspondence between equipment and processes, the root node of the product process tree is the last process of the product processing. When it is finished, it means that the whole product processing and assembly is completed. This paper proposes an integrated scheduling system for the same equipment process sequencing base on the Root-Subtree vertical and horizontal pre-scheduling. The directed graph of the product process tree is divided into several Root-Subtrees by the tree decomposition method, the vertical and horizontal characteristics of each Root-Subtree are evaluated respectively, and based on satisfying the constraint partial order relationship of the process tree, the Root-Subtrees are fused to obtain the final solution, and a reasonable scheduling scheme is sought to reduce the length of the equipment idle time. Due to the parallel relationship between Root-Subtrees, during the fusion scheduling process, there will be a problem that the Root-Subtree processes compete for the processing equipment horizontally. If no reasonable method is found, it will affect the progress of the entire production task. This paper uses the Root-Subtree pre-scheduling strategy to mine the vertical and horizontal characteristics of all Root-Subtrees, determines the priority of each Root-Subtree, and then sorts all Root-Subtree processes by the equipment grouping to solve the horizontal conflict problem of the Root-Subtree processes. On this basis, the dynamic start processing time strategy of the equipment process determines the start processing time of all Root-Subtree processes, the scheduling results of all Root-Subtree process are saved. Getting back the root node produces the actual product process scheduling sequence, solves the vertical conflict problem of the Root-Subtree processes, and finally obtains a reasonable scheduling scheme, which minimizes the product processing completion time. In this paper, the related definitions are described as follows:

Definition 1, Root-Subtree set: The product process tree is divided into several Root-Subtrees by using the tree decomposition method of the graph.

Definition 2, Root-Subtree process set: All the processes in the Root-Subtree.

Definition 3, Vertical pre-scheduling completion time: Through algorithm ① in the literature [

Definition 4, Horizontal pre-scheduling completion time: Through algorithm ② in the literature [

Definition 5, Aspect ratio of Root-Subtree: The ratio of the vertical pre-scheduling completion time to the time of horizontal pre-scheduling.

Definition 6, Virtual root node: If the in-degree of the root node (process) of the product process tree is not greater than 1, it is judged whether the in-degree of the immediately preceding process is greater than 1, and so on. The node with an in-degree greater than 1 is always found, and the node and its immediately following processes are merged to form a new root node.

Definition 7, Root-Subtree priority: During the fusion process, the parameters of the Root-Subtree process scheduling priority level are set from large to small according to the pre-scheduling completion time of the Root-Subtree, use

Definition 8, Equipment process pre-start time: In the analysis of the vertical and horizontal advantages of the Root-Subtree, the start processing time of each Root-Subtree process is obtained according to the selected pre-scheduling plan.

Definition 9, Non-schedulable process: The unprocessed process of the immediately preceding processes.

Definition 10, Scheduling process: The process group that has no immediately preceding processes or all immediately preceding processes have been scheduled.

Definition 11, Schedulable procedure set: The set of all schedulable processes.

Assuming that an existing complex product process tree model has

The processing time of each process on the processing equipment is determined and independent of other factors;

Each processing equipment can process only one working process at the same time; Once processed, it cannot be interrupted;

There is no equipment with the same function, which allows the processing equipment to wait idle;

There is parallel processing characteristic between Root-Subtrees, there is no restriction relation of the process tree, but there is competition constraint on processing equipment;

Each process must meet the constraints of the product process tree during the processing process, and it must be processed after all the immediately preceding processes have been processed;

After the processes are completed, the processing equipment becomes idle, other dispatchable processes can occupy the equipment.

Under the above conditions, the mathematical model of the integrated scheduling scheme is described as

Mathematical model,

According to the above equation, the description of the proposed algorithm in this paper further explains the main idea of the integrated scheduling algorithm for the same equipment process sequencing base on the Root-Subtree vertical and horizontal pre-scheduling, that is, the process tree is split and merged, it fully mines the internal characteristics of the product process tree, so that each process can be processed as early as possible, optimizes and minimizes the maximum completion time target, and improves the process processing compactness and the efficiency of the solution. The algorithm in this paper can be divided into three steps: The first step is to split the complex product process tree into several Root-Subtrees; The second step is to traverse each Root-Subtree to find the aspect ratio of the Root-Subtree, and analyze the vertical and horizontal characteristics of each Root-Subtree, so as to obtain the Root-Subtree priority, and solve the horizontal equipment competition relationship of the Root-Subtree process in the fusion process; The last step is to use the sorting strategy of the equipment process pre-start time and the dynamic start processing time strategy of the equipment process to perform fusion scheduling for all Root-Subtree processes. In this way, an optimal scheduling algorithm from the whole→part→whole execution process is completed, which considers the internal structure of the product process tree instead of only the overall structure.

There is a complex product process tree model with

After the splitting process tree operation is completed, then all Root-Subtrees are traversed respectively by the algorithm ① and the algorithm ② to obtain the vertical pre-scheduling completion time and the horizontal pre-scheduling completion time, and then to obtain the corresponding Root-Subtree aspect ratio. In contrast, this scheduling is not a real production scheduling but only a reference to analyzing the vertical and horizontal characteristics of the process tree. According to the aspect ratio results, it is analyzed that the Root-Subtree has vertical or horizontal advantages. The priority of each Root-Subtree is determined at the same time, to solve the horizontal competition relationship between Root-Subtrees on the processing equipment, and do an excellent job for the subsequent resolution of the process selection conflicts foreshadowing, the choice of which pre-scheduling plan is most suitable directly affects the optimization goal. Because each Root-Subtree is a branch of the product process tree, each Root-Subtree has the same optimization goal as the product process tree itself, which is to minimize the maximum completion time and enhance the scheduling compactness of the process. Using this criterion, the vertical and horizontal advantage of the Root-Subtree is determined by comparing the pre-scheduling completion time of the algorithm ① and the algorithm ② that is, it chooses a scheduling scheme with a short pre-scheduling completion time,

Situation 1, when

Situation 2, when

Situation 3, when

After the analysis of the vertical and horizontal characteristics of the Root-Subtree is completed, this paper uses the idea that the critical path of the product process tree has the greatest impact on the total processing time and that the critical path process is processed first. Compared with the critical path, the Root-Subtree is a branch of the product process tree, the critical path is only the longest path in the process tree. Therefore, using the Root-Subtree pre-scheduling completion time can better explain the lower limit of the product scheduling completion time, that is to say, the selected Root-Subtree with the longer pre-scheduling completion time has a more significant impact on the lower limit of the total product processing time, that is, the longer the pre-scheduling completion time, the higher the priority of the Root-Subtree in the process of fusion scheduling. If there are multiple Root-Subtrees with the same selected pre-scheduling completion time, it follows the principle that the more the number of the Root-Subtree processes, the higher is the priority for processing. The Root-Subtree priority strategy guarantees the flexibility of the Root-Subtree process scheduling, and takes priority to schedule the process of the Root-Subtree with high priority. The Root-Subtree priority strategy uses

To complete the production scheduling task of a complex product, it is necessary to merge the Root-Subtree process set with the parallel relationships [

Situation 1, if

Situation 2, if

Situation 3, if

The above analysis of the processing of two processes on only one equipment can be extended to the sorting processing of multiple equipment and multiple processes. Finally, the processing equipment is grouped to determine the scheduling processing sequence of the Root-Subtree processes.

In the scheduling process from the decomposition of the product process tree to the integration, the process scheduling needs to meet the constraints of both vertical and horizontal aspects: Horizontally, it must be satisfied that all the immediately preceding processes of the same equipment have been scheduled for processing; Vertically, it must be satisfied that all the immediately preceding processes of the process tree have been scheduled and processed. The sorting strategy of the equipment process pre-start time only solves the problem of the processing sequence of the same equipment process in the horizontal direction, so the dynamic start processing time strategy of the equipment process is proposed to solve the sequential constraint relationship of all processes in the process tree in the vertical direction [

Situation 1, the process

Situation 2, the process

Situation 3, the process

Situation 4, the process

Aiming at the complex product integrated scheduling problem with the internal characteristics of the product process tree, this paper uses the Root-Subtree pre-scheduling strategy to split the process tree into several Root-Subtrees and set the priority for each Root-Subtree. On this basis, through the Root-Subtree process set fusion conflict adjustment strategy, the equipment sequence is grouped as the Root-Subtree process set for sorting, and the start processing time of the process is adjusted to produce a substantial scheduling plan. The specific steps of the integrated scheduling algorithm for the same equipment process sequencing based on the Root-Subtree vertical and horizontal pre-scheduling are as follows:

Step 1: Perform data information processing on the complex product model to form a standardized integrated scheduling process tree model; Establish a schedulable process set, and add all schedulable processes in the product process tree to the schedulable process set.

Step 2: According to the simplified pretreatment principle and the process tree decomposition strategy, the root node is taken out and stored in independent storage space, and the process tree is split into several Root-Subtrees to form a Root-Subtree set {

Step 3: Perform pre-scheduling for each Root-Subtree through the algorithm ① and the algorithm ② and use the pre-scheduling results to obtain the aspect ratio

Step 4: Determine the priority of the Root-Subtree according to the selected pre-scheduling scheme, sort all Root-Subtrees from high to low using the bubble sorting method, and obtain the equipment process pre-start time at the same time. If there are multiple Root-Subtrees with the same pre-scheduled completion time, the priority is determined by the number of the Root-Subtree processes from more to less.

Step 5: All processes are grouped according to the corresponding equipment sequence

Step 6: Set the processing equipment stack {

Step 7: Judge whether all stacks are empty, if yes, go to Step 11, otherwise go to Step 8.

Step 8: Judge whether the stack-top process is a schedulable process through the schedulable process set. If no, the start processing time of the process is not updated, and the stack-top process is in the waiting stage without being ejected. If yes, the dynamic start processing time strategy of the equipment process is used to determine the start processing time

Step 9: Delete the scheduled processes from the schedulable process set, and add the new schedulable processes to the schedulable process set.

Step 10: Determine the initial start processing time

Step 11: Save the results of the essential scheduling sequence of the Root-Subtree process set.

Step 12: Based on the previous Step 11, the root node of the process tree extracted in Step 2 is added to the actual scheduling sequence, and finally the scheduling result of the entire production process tree is formed.

Step 13: Output the Gantt chart according to the scheduling result of the entire production process tree in Step 12.

The proposed algorithm flow chart in this paper is shown in

Assuming that the number of the complex product processes is

In simplifying and splitting the process tree, it is necessary to reverse the order to determine whether the root node of the process tree has a bifurcation. If so, the root node is directly taken out; if there is no bifurcation, the root node and the immediately preceding process are merged into a virtual root node, cyclically analyze whether there is a bifurcation in the immediately preceding process of the virtual root node. The worst case is that except for a leaf node, the remaining

The average number of the processes for each Root-Subtree is

The complexity of grouping all Root-Subtree processes with the corresponding processing equipment is

According to the dynamic start processing time strategy of the equipment process, it is necessary to establish and update the schedulable process set. The complexity of finding the current schedulable processes in the process tree for any equipment is

The proposed algorithm in this paper involves 4 main strategies, all strategies are connected in series, so the complexity of the algorithm is the sum of the complexity of each strategy. In summary, the complexity of the proposed algorithm is a quadratic polynomial

The proposed algorithm is a theoretical analysis process of the integrated scheduling, which is not based on a specific instance. To facilitate readers to understand the algorithm, and it reflects the universality of the algorithm and evaluates the performance of the algorithm, the following is an example for analysis and comparison. We suppose that a manufacturing company plans to complete an order and record it as the product H, consisting of 22 processes and requiring three processing equipment. The process tree is shown in

Firstly, according to the characteristics of the process tree structure of the product H, the process tree decomposition strategy is used to take out the root node “Root”, store it in an independent space, and split the process tree to form two Root-Subtrees, namely

Secondly, the algorithm ① and the algorithm ② are used to respectively traverse and pre-schedule each Root-Subtree of the product H, and then obtain the vertical pre-scheduling completion time

Root-Subtree name | Vertical pre-scheduling completion time | Horizontal pre-scheduling completion time | Aspect ratio | Selected pre-scheduling plan | Root-Subtree priority |
---|---|---|---|---|---|

Horizontal | |||||

Vertical |

Thirdly, it determines the equipment process pre-start time of all Root-Subtree processes through the selected pre-scheduling scheme in

Equipment name | Process name | The pre-start time of the equipment process | The Root-Subtree priority corresponding to the process | The same equipment process scheduling sequence |
---|---|---|---|---|

M1 | A11 | 0 | 1 | |

A10 | 10 | 3 | ||

A4 | 150 | 6 | ||

A2 | 100 | 5 | ||

B9 | 0 | 2 | ||

B6 | 40 | 4 | ||

B1 | 145 | 7 | ||

M2 | A9 | 10 | 3 | |

A8 | 0 | 1 | ||

A3 | 45 | 5 | ||

B10 | 0 | 2 | ||

B8 | 20 | 4 | ||

B4 | 75 | 6 | ||

B3 | 105 | 7 | ||

M3 | A7 | 100 | 5 | |

A6 | 0 | 1 | ||

A5 | 50 | 3 | ||

A1 | 170 | 7 | ||

B7 | 0 | 2 | ||

B5 | 55 | 4 | ||

B2 | 105 | 6 |

Fourth, the same equipment processes are sequentially pressed into the processing equipment stacks {

Equipment name | Root-Subtree process | Initial start processing time | Updated start processing time | Remarks |
---|---|---|---|---|

M1 | A11 | 0 | 0 | |

B9 | 10 | 10 | ||

A10 | 50 | 50 | ||

B6 | 90 | 90 | ||

A2 | 125 | 150 | Updated | |

A4 | 175 | 215 | Updated | |

B1 | 235 | 235 | ||

M2 | A8 | 0 | 0 | |

B10 | 10 | 10 | ||

A9 | 30 | 30 | ||

B8 | 50 | 50 | ||

A3 | 85 | 85 | ||

B4 | 115 | 125 | Updated | |

B3 | 155 | 165 | Updated | |

M3 | A6 | 0 | 0 | |

B7 | 45 | 45 | ||

A5 | 100 | 100 | ||

B5 | 150 | 150 | ||

A7 | 165 | 165 | ||

B2 | 215 | 215 | ||

A1 | 235 | 235 |

Finally, retrieving the root node “Root” to enter the scheduling sequence to form the actual scheduling plan of the product H and complete the product production scheduling task. The Gantt chart of the scheduling product H using the proposed algorithm is shown in

The following uses DCPM, DJSSA and ISA-CPTS to schedule the product H and analyze the scheduling results to illustrate the superiority of the proposed algorithm further. The Gantt chart of DCPM for the overall scheduling of the process tree H mainly in the vertical direction is shown in

In order to verify the effectiveness of ISA-SEPS and its adaptability to products with different structures, this paper randomly selects 50 product models with different structures, the selected product instances were used for scheduling analysis using DCPM, DJSSA, ISA-CPTS and ISA-SEPS. The specifications of all process tree model parameters are as follows: the number of the process is [20,50], the number of immediately preceding processes of each process is [0,5], the number of the Root-Subtree is [2,5], and the number of the processing equipment is [3,5], the process processing time is [10,60].

In the field of the integrated scheduling, this paper proposes for the first time an integrated scheduling algorithm for the same equipment process sequencing base on the Root-Subtree vertical and horizontal pre-scheduling. According to the decomposition principle of the product process tree, the process tree is split into several Root-Subtrees, and the vertical and horizontal characteristics of each Root-Subtree are excavated through the pre-scheduling method. On this basis, the sorting strategy of the equipment process pre-start time is used to sort the Root-Subtree process set with the processing equipment sequence as the grouping to solve the horizontal constraint relationship with the same equipment, at the same time, the dynamic start processing time strategy of the equipment process is used to solve the constraint relationship of the vertical process tree, thereby completing an optimized scheduling process considering the internal structure of the product process tree. The test results show that the proposed algorithm is better than the traditional integrated scheduling algorithm for scheduling processes as a whole. It has the following characteristics: (1) It has better adaptability to product instances of different structures; (2) It excavates the vertical and horizontal characteristics of the product process tree, enhances the compactness of equipment process scheduling, and conforms to the principle of early processing and early termination; (3) The scheduling process has better flexibility and reduces the idle time of the processing equipment. In summary, the proposed algorithm ideas can provide a reasonable scheduling scheme for the product process trees with different structural, and expand a new direction for in-depth study of integrated scheduling problems. In addition, the proposed algorithm can be further applied to the field of the dynamic integrated scheduling, so the algorithm has certain theoretical and practical significance.