CN100394417C - A rule scheduling method with state performance feedback and its scheduling system - Google Patents

Abstract

A rule scheduling method with state performance feedback and its scheduling system can reduce workpiece delay time and improve equipment utilization. On the basis of the existing rule scheduling, introduce equipment demand, equipment demand trend, and workpiece average slack time as system operation status feedback, and introduce equipment processing time, workpiece processing time, workpiece waiting time, and workpiece advance/delay time as system operation Performance feedback to form a feedback scheduling rule. The scheduling objective function is introduced and an iterative optimal scheduling method is proposed. The extended process sequence planning tree is defined and used to represent the process sequence planning flexibility of the workpiece and the processing path flexibility of the workpiece, which expands the decision space of the scheduling system and is conducive to forming a better scheduling scheme. The dispatching system is composed of central controller, loading and unloading station, system buffer station, automatic track trolley and production equipment. At each decision-making moment, the central controller calculates the priority of idle equipment and workpieces to be processed according to the feedback scheduling rules, and selects idle equipment with high priority to process workpieces with high priority.

Description

A rule scheduling method with state performance feedback and its scheduling system

technical field

The invention belongs to the technical field of production scheduling and management, and in particular relates to a rule scheduling method with state performance feedback and a scheduling system thereof.

Background technique

According to the optimal production technology, the production capacity of the bottleneck equipment determines the production capacity of the whole system and determines the operating performance of the system. For a production system that has been put into production, the processing capacity of its equipment is fixed. Only by using an appropriate scheduling system and scheduling method can the logistics in the production system be changed and the demand of the workpiece on the production capacity of the equipment be adjusted to reduce or eliminate the bottleneck of the system. effect to improve the performance of the entire system.

The scale of the production scheduling problem increases with the increase of the number of production equipment, the variety of workpieces, the number of workpieces, and the number of process planning, and the computational complexity of the scheduling problem increases exponentially with the increase of the scale of the scheduling problem. Therefore, for large-scale production systems with complex logistics conditions, the rule-based scheduling method is generally used to complete real-time scheduling and control of the production system.

Scheduling rules are a sorting method used to determine the competing priorities of production equipment and workpieces to be processed in a production system. Whenever the equipment is idle, or a new blank enters the system, or a process of the workpiece is processed, it enters a decision-making moment. At this time, the scheduling system determines which process of which workpiece is allocated to which equipment according to the scheduling rules. In order to implement scheduling and control on the production process, this scheduling method is called rule-based scheduling, or rule scheduling for short. Under the control of scheduling rules, the process queue assigned to each equipment in the production system and the start and end time of each process are called the solution of rule scheduling.

At a certain decision-making moment, if there are multiple processing equipment in the system that can complete the processing of the same workpiece to be processed, the production equipment with the highest priority is first selected to complete the processing task. Similarly, at a certain decision-making moment, if there are multiple workpieces waiting to be processed in the system, the workpiece with the highest priority is first selected for processing. Through scheduling rules, it is possible to prioritize the production equipment and workpieces to be processed in the production system. Different scheduling rules can produce different sorting results, which can change the flow process of workpieces in the production system and achieve the purpose of controlling the operation process and performance of the production system.

Generally speaking, regular scheduling has the advantage of being simple and easy to implement, and has dynamic scheduling capabilities, which is especially suitable for the scheduling of complex production systems with many random interference factors in the production process. See: A. Baker, A survey of factory control algorithms that can be implemented in a multi-agent heterarchy: dispatching, scheduling, and pull, J. of Manufacturing Systems, 1998, 17(4): 297-320.

The disadvantage of regular scheduling is that it cannot guarantee the optimal solution, or even a satisfactory solution in many cases. The main reason is that regular scheduling lacks scheduling objective functions and optimization control methods for objective functions. Once the scheduling rules are determined, the sorting method of production equipment and workpieces to be processed will be determined accordingly. Therefore, rule-based scheduling cannot adjust the sorting method of production equipment and workpieces to be processed in real time according to the system's operating status and performance, and cannot be based on the actual operating conditions of the system. Optimizing the production process, especially the production process cannot be optimized for the current scheduling objective function. On the other hand, the existing rule scheduling does not make full use of the flexibility of the process sequence planning of the workpiece and the flexibility of the processing path of the workpiece, so the decision-making space is small, and there are few alternative scheduling schemes, which is not conducive to forming a better scheduling scheme.

Contents of the invention

The purpose of the present invention is to provide a rule scheduling method with state performance feedback that can overcome the above-mentioned defects. This method can adjust the production equipment and the sorting method of the workpieces to be processed in real time according to the system operating status and operating performance, and can adjust according to the actual operating conditions of the system. Optimizing the production process, especially optimizing the production process for the current scheduling objective function; the present invention can also make full use of the flexibility of the process sequence planning of the workpiece and the flexibility of the processing path of the workpiece in the scheduling process to form a better scheduling scheme; the present invention also A scheduling system implementing the method is provided.

A rule scheduling method with state performance feedback provided by the present invention comprises the following steps in turn:

(1) Formulate the scheduling objective function Obj according to the production task;

(2) Specify scheduling rules r _{M, k} , r _J, i for production equipment k and workpiece i respectively, k=1, 2,..., k, i=1, 2,... ., N;

(3) Construct iterative feedback scheduling rules r _cM,k , r _cJ,i :

r _{cM, k} = r _{M, k} +p _k1 R _k +p _k2 D _k +q _k T _k

r _{cJ, i} = r _{J, i} + p _i t _{s, i} + q _i1 t _{p, i} /t _{w, i} -q _i2 t _{d, i}

Among them, p _k1 and p _k2 are the state feedback coefficients of equipment k, q _k is the performance feedback coefficient of equipment k, p _i is the state feedback coefficient of workpiece i, q _i1 and q _i2 are the performance feedback coefficients of workpiece i; R _k is The demand of equipment k, D _k is the demand trend of equipment k, T _k is the processing time of equipment k; t _{s, i} is the average relaxation time of workpiece i, t _{p, i} is the processing time of workpiece i, t _{w , i} is the waiting time of job i, t _{d, i} is the delay time of job i; where,

Equipment demand R _k (t) is calculated by the following formula:

R _k (t)=|{J _i (t)}|, where J _i (t) represents the workpieces that can be processed on the production equipment at time t, {} represents the set, and || represents the basis of the set;

The equipment demand trend D _k (t) is:

${\mathrm{D.}}_k\left(t\right)=\frac{d}{\mathrm{dt}}{f}_k\left(t\right),$ In the formula, F _k (t) is a continuous derivable function generated by fitting according to R _k (t);

The average relaxation time t _{s of workpiece i, i} is:

$t_{\mathrm{the\; s},i}=\frac{d_i-t-t_{r,i}}{{\mathrm{no}}_{r,i}},$ In the formula, t is the current moment, d _i is the planned delivery time of workpiece i, t _{r, i} is the expected value of the total processing time required for the unprocessed process of workpiece i, n _{r, i} is the number of unprocessed processes of workpiece i ;

(4) Set iteration control parameters and iteration initial value:

Let the current number of iterations j=1, let the maximum number of iterations be N _max ;

Let S _opt be empty, and let Obj(S _opt ) be a sufficiently large number;

Let T ⁽⁰⁾ _k = 0, t ⁽⁰⁾ _{p, i} = t ⁽⁰⁾ _{w, i} = t ⁽⁰⁾ _{d, i} = 0;

(5) All equipment is set to idle state, and all workpieces are set to initial state;

(6) Assign workpieces to be processed to idle equipment according to the following steps:

(6.1) If there is no idle equipment or no workpiece to be processed, go to step (7);

(6.2) Utilize the following formula to calculate the priority of all idle devices;

r ^(j) _{cM, k} = r _{M, k} +p _k1 R ^(j) _k +p _k2 D ^(j) _k +q _k T ^(j-1) _k

(6.3) Complete each idle device in order of priority from high to low:

(6.3.1) Determine the set L of workpieces to be processed that can be processed on it, if L is an empty set, then take the equipment with the second priority as the idle equipment, and repeat this step;

(6.3.2) Calculate the priority of workpieces to be processed in L according to the following formula:

r ^(j) _{cJ, i} = r _{J, i} + p _i t ^(j) _{s, i} + q _i1 t ^(j-1) _{p, i} /t ^(j-1) _{w, i} -q _i2 t ^{(j -1)} _d,i

(6.3.3) Take the workpiece with the highest priority and process it for the idle device;

(6.3.4) judge whether there is also idle equipment, if so, enter step (6.3.1), otherwise, enter step (7);

(7) If all workpieces are processed, turn to step (8); otherwise, advance one unit of time, turn to step (6);

(8) At the end of this iteration, calculate the operating performance of this iteration system, and calculate the scheduling objective function value Obj(S ^(j) ), where j is the current iteration number. If Obj(S ^(j) )<Obj(S _opt ), then S _opt =S ^(j) ;

(9) Make j=j+1, if j≤N _max , go to step (5), otherwise go to step (10);

(10) Stop the iteration and output the optimal scheduling result, which can be directly used for the scheduling of the same production task.

The scheduling system that realizes the above method includes a central controller, a loading and unloading station, a system buffer station, an automatic rail trolley and processing equipment, and is characterized in that:

The processing equipment is equipped with an equipment controller;

The blank enters the system buffer station through the loading and unloading station to wait for processing, and the finished product exits the system through the loading and unloading station;

The system buffer station is used to temporarily store workpieces, and it is equipped with a buffer station controller; the system buffer station is equipped with physical stations, and each physical station is equipped with a pallet, on which workpieces to be processed are clamped; the buffer station controller records the objects The relative position of the engineering station relative to the absolute origin of the system buffer station, and also records the state of the physical station;

The automatic track trolley can move between the loading and unloading station, the system buffer station and the processing equipment, and it is equipped with an automatic pallet picking/putting mechanism;

The loading and unloading station, the system buffer station and the processing equipment are equipped with pallet exchange stations, each pallet exchange station is equipped with a position indicating element, and the position detection element installed on the automatic track trolley can detect each pallet exchange station , so as to realize the precise positioning of the automatic track trolley;

The central controller is respectively connected with the display device of the loading and unloading station, the input device of the loading and unloading station, the automatic track trolley, the controller of the system buffer station and the controller of the production equipment to control the work of each component.

In the scheduling method and system involved in the present invention, on the basis of the existing rule scheduling, the feedback of system running status and running performance is added, and the sorting method of production equipment and workpieces to be processed is adjusted in real time through the feedback mechanism to control the production system. Logistics, which matches the demand of production equipment for workpieces to be processed with the processing capacity of production equipment, reduces or eliminates the bottleneck equipment effect of the system, optimizes production operation, and thus overcomes the problems existing in the existing rule scheduling method. At the same time, the present invention also provides a technical means to realize the production target of the production manager and optimize the specified operating performance, that is, to establish a scheduling objective function according to the system performance index that the production manager is most concerned about at present, and to use the feedback scheduling rule on the basis of The iterative optimization method gradually improves the operating performance of the system in the sense of the scheduling objective function. This point is of great significance in actual production scheduling. The production manager can formulate the scheduling objective function of the current production task according to the performance index he cares most about. The scheduling method and scheduling system can maximize the performance of related systems.

The scheduling method proposed by the present invention tests three scheduling objective functions respectively: Obj1 (minimum workpiece delay time), Obj2 (maximum equipment utilization rate) and Obj3 (minimum workpiece delay time and maximum equipment utilization rate), and compares with existing Compared with the regular scheduling method, the statistical analysis of the experimental results shows that: when the scheduling goal is to minimize the delay time of the workpiece, compared with the existing rule scheduling method, at the 0.001 significant level (t-test), this method can reduce the delay time of the workpiece 96.7% or more, while the utilization rate of equipment is slightly improved (<2%). When taking the maximum equipment utilization rate as the scheduling objective function, compared with the existing regular scheduling method, at a significant level of 0.001 (t-test), this method can increase the equipment utilization rate by more than 3.2%, while the workpiece delay time is only reduced by 2.2% . When the scheduling objective function is the minimum workpiece delay time and the maximum equipment utilization rate, compared with the existing regular scheduling method, at a significant level of 0.001 (t-test), this method can reduce the workpiece delay time by more than 93.5%, and improve equipment utilization rate above 4%. The test results of the present invention show that: for a given scheduling rule, the performance of the scheduling method is generally better than that of the existing rule scheduling method.

Description of drawings

Figure 1 is an extended process sequence planning tree;

Figure 2 is a schematic diagram of the principle of rule scheduling with state performance feedback;

Fig. 3 is a schematic diagram of the decision-making process at the decision-making moment of the scheduling method of the present invention;

Fig. 4 is a kind of equipment composition and layout model of production system;

Fig. 5 is a schematic diagram of the communication connection of the dispatching system.

Detailed ways

In the production system, the production equipment M has a certain processing capacity and can complete certain processing operations. The workpiece is the object to be processed, and the processing of the workpiece can be divided into several different processing operations. Each processing operation is called a process Op of the workpiece. In the process of workpiece processing, some processes must be completed before other processes, that is, there are certain requirements for the processing sequence of the processes. A feasible processing sequence of the workpiece is called a process sequence planning of the workpiece. When there is more than one feasible process sequence plan for a workpiece, it is said that the workpiece has multiple process sequence plans, or the workpiece has process sequence planning flexibility.

Each process of the workpiece can be processed on one or more production equipment, but at the same time, a workpiece can only be processed on one equipment at most. When a process of a workpiece is processed on different production equipment, the required processing time is not the same. In order to complete the processing of the workpiece, each process of the workpiece must be processed in sequence according to the process sequence planning requirements of the workpiece. For this reason, the workpiece must visit the production equipment in a certain order and be processed on the equipment. The order in which workpieces access production equipment in the system is called the processing path of workpieces. If a process of a workpiece can be processed on multiple production equipment, there are multiple processing paths for the workpiece in the production system, which is called the flexibility of the processing path of the workpiece.

In addition, due to the requirements of the production process, different workpieces may have different processing paths, and the same workpiece may visit the same production equipment multiple times.

From the above discussion, it can be seen that the present invention does not impose any additional restrictions on the production process, and is the most general production process. Therefore, the scheduling method and scheduling system proposed by the present invention have wide adaptability, and are suitable for small batches of products and product varieties. The scheduling of production systems with multiple and complex logistics processes, such as the scheduling of flexible manufacturing systems, is also suitable for the scheduling of mass production systems, flow production systems, single-piece job production systems, and reentrant production systems.

The dispatching method and dispatching system proposed by the present invention aim at rationally arranging the processing of the workpiece process and improving the operating performance of the production system according to the extended process sequence planning tree, with the aid of status and performance feedback, and with the dispatching objective function as the optimization index.

The present invention proposes an extended process sequence planning tree (Fig. 1) model and is used to represent the flexibility of the process sequence planning of the workpiece and the flexibility of the processing path of the workpiece. In the figure, a node represents a process, represented by Op, and the elements in {} The number represents the number of equipment in the production system, and the sequence of numbers in {} indicates the relationship between the process and the equipment in the production system, where "-1" indicates that the process cannot be processed on this equipment, and a positive number indicates that the process can be processed on this equipment Processing and the magnitude of the value indicates the required processing time. For example, Op1{10, -1, -1, -1, 14, -1, -1, -1, -1, -1} means that there are 10 devices in the production system, and the operation Op1 can be performed on devices 1 and 5 processing, and the required processing time is 10 and 14 unit times respectively.

A path from the root node to the leaf node is called a process sequence planning scheme of the workpiece. For the workpiece shown in Figure 1, the workpiece consists of 4 processes, namely Op1, Op2, Op3 and Op4. There are 3 feasible process sequence plans for the processing of the workpiece, namely:

(1) Op1→Op3→Op4→Op2;

(2) Op1→Op4→Op3→Op2;

(3) Op1→Op4→Op2→Op3.

Among them, the process sequence planning (1) means: after Op1 is processed, Op3 can be processed; after Op3 is processed, Op4 can be processed; after Op4 is processed, Op2 can be processed; and so on.

Since both Op1 and Op4 can be processed on two devices, for any process sequence planning, the workpiece has 4 feasible processing paths in the production system, for example, for process sequence planning (1), the workpiece’s 4 The feasible processing paths are:

(1) M1→M2→M4→M3

(2) M1→M2→M8→M3

(3) M5→M2→M4→M3

(4) M5→M2→M8→M3

The extended process sequence planning tree can represent the process sequence planning flexibility of the workpiece and the processing path flexibility of the workpiece, so that the flexibility of the system and the workpiece can be effectively used in the rule scheduling process to form a better scheduling scheme.

Rule scheduling aims at arranging a suitable process for each idle device at every decision moment. For example, for the workpiece shown in Figure 1, after Op1 is processed, the next processing process can choose Op3 or Op4. Op3 can be processed on equipment M2, and the required processing time is 30 unit times; while Op4 can be processed on equipment M4 and M8, and the required processing time is 24 and 31 unit times respectively. Which process to choose and which equipment to process the process are determined by the scheduling rules. In the present invention, scheduling rules with status and performance feedback are used to determine the priorities of equipment and workpieces, and equipment with high priority is selected to process workpieces with high priority, thereby completing production scheduling.

During the operation of the production system, the variables that describe the operating properties of the equipment and workpieces in the system are called system state variables. For the state variable, we mainly care about its changing process in the whole scheduling cycle, not only its value at the end of scheduling. The operating status of the production system is divided into equipment status and workpiece status.

The quantity that reflects the demand of each production equipment for the workpiece to be processed is called the equipment state variable. Equipment status variables include equipment demand quantities and equipment demand trends.

At time t, the demand of all workpieces to be processed on the production equipment Mk in the production system is called the demand of equipment k, denoted as R _k (t). The equipment demand is calculated by the following formula:

R _k (t)＝|{J _i (t)}| (Formula 1a)

Among them: J _i (t) represents the workpiece that can be processed on the production equipment Mk at time t, {} represents the set, and || represents the basis of the set.

For the workpiece shown in Figure 1, suppose the workpiece has entered the system and has not been processed at time t, then the workpiece can be processed on equipment M1 and M5, and there is a demand for these two equipments; if the process Op1 of the workpiece has been processed, then Op3 or Op4 can be selected for the next process, where Op3 can be processed on equipment M2, and Op4 can be processed on equipment M4 or M8, so the workpiece has requirements for equipment M2, M4 and M8. According to the extended process sequence planning tree, the equipment requirements of all workpieces to be processed can be calculated, so that the demand for each idle equipment can be calculated.

Equipment requirements and the queue length of workpieces to be processed in front of the equipment are different concepts. The queue length indicates the number of workpieces that have been assigned to the equipment but has not yet been processed; while the equipment demand indicates the potential demand of the equipment by the workpieces in the system. Due to the flexibility of process sequence planning and processing path of workpieces, these workpieces may be allocated to this equipment for processing, or may be allocated to other equipment for processing.

The changing trend of the demand for equipment Mk for workpieces to be processed in a period of time in the future is called the demand trend of equipment k, denoted as D _k (t). Assume that the continuous derivable function F _k (t) can be used to fit the equipment demand R _k (t), then the equipment demand trend is:

${\mathrm{D.}}_k\left(t\right)=\frac{d}{\mathrm{dt}}{f}_k\left(t\right)$ (Formula 1b)

In the production system, the quantity that reflects the attributes of the processed workpiece is called the workpiece state variable. The workpiece state variables mainly include the average relaxation time (t _{s, i} ) of the workpiece i, and its calculation method is as follows:

$t_{\mathrm{the\; s},i}=\frac{d_i-t-t_{r,i}}{{\mathrm{no}}_{r,i}}$ (Formula 1c)

In the formula, t is the current moment, d _i is the planned delivery time of workpiece i, t _{r, i} is the expected value of the total processing time required for the unprocessed process of workpiece i, n _{r, i} is the number of unprocessed processes of workpiece i .

The system state variable defined by the present invention is:

$x=\left[\begin{array}{c}R\\ \mathrm{D.}\\ t_{\mathrm{the\; s}}\end{array}\right]$ (Formula 1d)

Among them, R=[R ₁ , R ₂ ,..., R _K ] ^T , D=[D ₁ , D ₂ ,..., D _K ] ^T , t _s =[t _{s , 1} , t _{s, 2} ,..., t _{s, N} ] ^T , K is the number of production equipment in the system, and N is the number of workpieces in the system.

Production system operation performance index is an evaluation standard to measure the operation of the production system. For running performance, the main concern is its value at the end of the production cycle, rather than its change process throughout the production cycle. Production system operation performance indicators can be divided into two categories: equipment-related indicators and workpiece-related indicators, and workpiece-related indicators can be further divided into workpiece completion indicators and workpiece flow indicators.

The index to evaluate the performance of equipment in the production system is called the equipment correlation index, and the key factor to measure the equipment correlation index is the processing time of the equipment in the entire scheduling cycle. The equipment correlation index mainly includes the average utilization rate (U) and equipment load unbalance rate (B) of the equipment in the system, and its calculation method is as follows:

$u=\left(\frac{1}{\mathrm{KT}}\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{T}_k\right)\&times;100\%$ (Formula 2a)

$B=\left(\frac{1}{K}\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}\right|T_k-\frac{1}{K}\underset{j=1}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\left|\right)\&times;100\%$ (Formula 2b)

Among them, T is the scheduling period, T _j and T _k are the processing time of equipment j and k respectively (including the loading and unloading time of the processed workpiece, the same below).

Equipment load unbalance rate is an important indicator to measure the relative load of equipment. When the load unbalance rate is large, it indicates that the equipment load distribution in the production system is uneven, there are bottleneck equipment, and the system productivity is limited by the productivity of these bottleneck equipment; when the load unbalance rate is small, it indicates that the equipment load distribution in the system is uniform.

The index to evaluate the completion of the processed workpiece in the production system is called the workpiece completion index, and the key factor to measure the completion index of the workpiece is the delay time of the workpiece. Workpiece completion indicators are also called workpiece relative indicators. Their evaluation criteria are relative to the planned delivery time of workpieces. These indicators reflect the system's ability to respond to market changes and are important indicators to measure the competitiveness of enterprises. Workpiece completion indicators mainly include the average workpiece delay time (t _d ), the maximum workpiece delay time (t _{d, max} ) and the number of delayed workpieces (N _d ), and their calculation methods are as follows:

$t_d=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}\left(\max \right\{t_{d,i},0\left\}\right)$ (Formula 2c)

t _{d, mam} = max{max{t _{d, i} }, 0} (formula 2d)

N _d =|{i,t _d,i >0}| (Formula 2e)

Where: t _{d, i} is the lead time/delay time of workpiece i, which is defined as:

t _d,i =t _c,i -d _i

Among them: t _{c, i} is the completion time of workpiece i, and d _i is the planned delivery time of workpiece i.

The index to evaluate the flow state of the processed workpiece in the production system is called the workpiece flow index, and the key factors to measure the workpiece flow index are the workpiece processing time and workpiece waiting time. Workpiece flow indicators are also called workpiece absolute indicators, which reflect the absolute status of workpieces during processing. These indicators reflect the production efficiency of the production system and are related to production costs. The workpiece flow index mainly includes the average processing time of the workpiece (t _p ), the average waiting time of the workpiece (t _w ) and the maximum passing time of the workpiece (t _{t, max} ), and its calculation method is as follows:

$t_p=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{p,i}$ (Formula 2f)

$t_w=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{w,i}$ (Formula 2g)

t _{t, max} = max {t _{p, i} +t _{w, i} } (formula 2h)

Among them: N is the number of workpieces in the system, t _{p, i} is the processing time of workpiece i (including loading and unloading time, the same below), t _{w, i} is the waiting time of workpiece i.

Table 1 gives the typical evaluation indicators of production system operation performance and the key factors to measure these indicators. It can be seen from Table 1 that although there are many indicators to evaluate the operation performance of the production system, the key factors to measure the operation performance are: in the entire production cycle, the processing time of the production equipment (T _k ), the processing time of the workpiece (t _{p, i} ), workpiece waiting time (t _{w, i} ) and workpiece delay time (t _{d, i} ). In the scheduling method involved in the present invention, the performance feedback variable is defined as:

y=[T ₁ ... T _K t _{p, 1} ... t _{p, N} t _{w, 1} ... t _{w, N} t _{d, 1} ... t _{d, N} ] ^T (Formula 2i)

Table 1

The dispatching method proposed by the present invention introduces state and performance feedback on the basis of existing rule dispatching, thereby constituting a feedback dispatching rule, and its principle is shown in FIG. 2 . In the figure, _Task is the task to be completed by the production system 1, r is the scheduling rule vector specified by the production manager H, r _c is the feedback scheduling rule vector, Obj is the scheduling target set by the production manager H, and x is the system running state vector, y is the system operation performance feedback vector.

Let P and Q be the state and performance feedback coefficient matrices respectively, then the matrix expression of the feedback scheduling rule is:

r _c ＝r+Px+Qy (Formula 3)

The above formula gives the general matrix expression form of the feedback scheduling rule of the rule scheduling method with state performance feedback. According to the specific production system, different system states and operating performance can be selected to form corresponding feedback scheduling rules.

In the scheduling algorithm of the present invention, for equipment k, the scheduling rule with status and performance feedback is:

r _{cM, k} = r _{M, k} + p _k1 R _k + p _k2 D _k + q _k T _k (Formula 3a)

For the workpiece i to be processed, the scheduling rule with status and performance feedback is:

r _{cJ, i} = r _{J, i} + p _i t _{s, i} + q _i1 t _{p, i} /t _{w, i} -q _i2 t _{d, i} (Formula 3b)

Among them, r _{M, k} , r _{J, i} represent the scheduling rules specified by the production manager for equipment k and the scheduling rules specified for workpiece i respectively, p _k1 and p _k2 are the state feedback coefficients of equipment k, and q _k is the status feedback coefficient of equipment k , p _i is the state feedback coefficient of workpiece i, q _i1 and q _i2 are the performance feedback coefficients of workpiece i. A simplification of the above formula is as follows: specify a scheduling rule r _M for all equipment, specify a scheduling rule r _J for all workpieces, and then form a feedback scheduling rule on this basis. A more simplified case is to specify a scheduling rule for all equipment and workpieces.

Order: r _c = [r _{cM, 1} r _{cM, 2} ... r _{cM, K} r _{cJ, 1} r _{cJ, 2} ... r _{cJ, N} ] ^T

r=[r _{M, 1} r _{M, 2} ... r _{M, K} r _{J, 1} r _{J, 2} ... r _{J, N} ] ^T

x＝[R ₁ R ₂ ... R _K D ₁ D ₂ ... D _K t _{s, 1} t _{s, 2} ... t _{s, N} ] ^T

y=[T ₁ T ₂ ... T _K u ₁ u ₂ ... u _N t _{d, 1} t _{d, 2} ... t _{d, N} ] ^T , (where u _i = t _{p, i} /t _{w, i} )

Then (Formula 3a) and (Formula 3b) can be expressed as a matrix expression in the form of (Formula 3).

The feedback coefficients in Equation 3a and Equation 3b reflect the relative influence of the corresponding state and performance on the sorting rules. In general, the state and performance feedback coefficients can be simply set to 1. If it is necessary to strengthen the optimization of a certain operating performance, then The corresponding state and performance feedback coefficients can be increased and decreased. For example: assuming that a production task has very strict requirements on delivery time, q _2i in (Formula 3b) can be increased. In particular, when the feedback coefficient is 0, it means that the scheduling rule has nothing to do with the corresponding feedback variable.

At any decision-making moment, calculate the priority of the idle equipment according to (Formula 3a), and determine the workpieces to be processed on a certain idle equipment in order of priority from high to low, and calculate according to (Formula 3b) The priority of the workpiece to be processed, assign the workpiece with the highest priority to be processed by the equipment. The regular scheduling of the production system is to apply this step repeatedly, and at each decision-making moment, assign the appropriate workpiece to the appropriate equipment for processing until all workpieces are processed.

In priority calculation, the smaller the value of r _{cM, k} , the higher the priority of the corresponding equipment; the smaller the value of r _{cJ, i} , the higher the priority of the corresponding workpiece.

When calculating the priority of idle equipment and workpieces according to (Formula 3a) and (Formula 3b), since state feedback is introduced into the priority calculation formula, the priority calculation method of equipment and workpieces is not fixed. The calculation of the level is related to the operating state of the system, thus providing a technical means to adjust the priority of equipment and workpieces in real time according to the operating state of the system. At the same time, (Equation 3a) and (Equation 3b) also introduce system operating performance feedback, thus providing an iterative optimization method to adjust equipment and workpiece priority ranking methods according to the last operating performance of the system.

The control process of feedback rule scheduling on logistics in the production system is as follows: when the utilization rate of equipment k is low, T _k decreases, through feedback control, r _k decreases, and the priority of equipment k increases, so equipment k can obtain more workpieces , to increase its utilization. Similarly, when job i is expected to fail to be completed on schedule, t _{d, i} increases, and through feedback control, r _i decreases, and the priority of job i rises so that it can be processed in advance.

The system operation performance is an objective evaluation of the system operation, while the scheduling target is the pursuit of the production manager for the specific operation performance of the system. Scheduling objectives reflect the subjective wishes of production managers. Therefore, different production managers often have different scheduling objectives. Even for the same production manager, their scheduling objectives will change with changes in production tasks and production cycles. The scheduling objective reflects the system performance that the production manager is most concerned about at present. Therefore, the optimization of the production process should be carried out around the scheduling objective, that is, focusing on optimizing the system operating performance included in the scheduling objective. In actual production scheduling, the scheduling target can be expressed as a function of the system's operating performance and system operating status. Common scheduling objective functions include: the shortest production cycle, the shortest workpiece delay time, the highest equipment utilization rate, the smallest equipment load imbalance rate, etc. The scheduling objective function may also be the weighted sum of the above indicators.

The dispatching system and dispatching method proposed by the present invention provide a technical means to optimize the system operating performance in the dispatching objective function by adjusting the status and performance feedback coefficients in the feedback dispatching rules. The specific adjustment method is as follows: for the dispatching objective function The status and performance variables appearing in , then increase the feedback coefficients of these variables in the feedback scheduling rules. For example, when the scheduling objective function is the minimum workpiece delay time, or the scheduling objective function includes the workpiece delay time variable, then increase the feedback coefficient q _i2 in (Equation 3b) to make the feedback scheduling rule more sensitive to workpiece delay time.

In addition, when there is a scheduling objective function in the system, the scheduling method also adopts the following iterative optimization strategy to continuously optimize the production process.

r ^(j) _{cM, k} = r _{M, k} + p _k1 R ^(j) _k + p _k2 D ^(j) _k + q _k T ^(j-1) _k (Equation 4a)

r ^(j) _{cJ, i} = r _{J, i} + p _i t ^(j) _{s, i} + q _i1 t ^(j-1) _{p, i} /t ^(j-1) _{w, i} -q _i2 t ^{(j -1)} _{d, i} (Equation 4b)

S _opt ＝S ^(j) , when Obj(S ^(j) )＜Obj(S _opt ) (Formula 4c)

In the formula, the superscript j represents the number of iterations. r ^(j) _cM,k , r ^(j) _{cJ, i} respectively represent the feedback scheduling rules adopted by equipment k and workpiece i in the j-th iteration, that is, at each decision point, according to this scheduling rule, determine the priority. S ^(j) is the system solution at the jth iteration obtained according to the feedback scheduling rule r ^(j) _{cM, k} and r ^(j) _{cJ, i} , that is, the process queue completed on each production equipment and each process The start time and end time of the operation. S _opt is the optimal solution of the system. Obj is the scheduling objective function, and Obj(S) represents the objective function value of the system when the system scheduling solution is S. In the first iteration, S _opt is defined as empty, and the initial value of Obj(S _opt ) is defined as a sufficiently large positive number, such as 1000, 10000 and so on. Here, in the jth iterative scheduling, the method of calculating the priority of equipment and workpieces is not only related to the current state of the system, but also related to the operating performance of the (j-1)th time system, thus providing a method based on the last operation of the system Iterative optimization means of performance adjustment equipment and workpiece priority calculation method.

Assuming that the production system has K sets of processing equipment, it is now necessary to complete a production task containing N workpieces, and the delivery time d _i of each workpiece in this production task is known, and the extended process sequence of each workpiece in this production task is known planning tree, the specific steps of applying the present invention to realize production scheduling are as follows:

1. Formulate the scheduling objective function Obj according to the production task

If the delivery time of the production task is tight or the late delivery will be punished, formula 2c, formula 2d or formula 2e can be selected as the scheduling target, namely:

$\mathrm{Obj}=\min t_d=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}\left(\max \right\{t_{d,i},0\left\}\right),$ or

Obj = mint _{d, max} = min(max{max{t _{d, i} }, 0}), or

Obj＝minN _d ＝min|{i,t _d,i >0}|

If the delivery time of the production task is loose, formula 2a or formula 2b can be selected as the scheduling target, namely:

$\mathrm{Obj}=\max u=\max \left(\frac{1}{\mathrm{KT}}\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{T}_k\right)\&times;100\%,$ or

$\mathrm{<span\; class="notranslate">\; Obj</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; K</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; K</span>}}\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; 100</span>}<span\; class="notranslate">\; \%</span>$

Formula 2f, formula 2g or formula 2h can also be selected as the scheduling target, namely:

$\mathrm{Obj}=\min t_p=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{p,i},$ or

$\mathrm{Obj}=\min t_w=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{w,i},$ or

Obj = mint _{t, max} = min(max {t _{p, i} + t _{w, i} })

The scheduling objective function can be a weighted sum of several performances in order to obtain better comprehensive performance. For example, when it is required to guarantee the delivery time of the workpiece and improve the utilization rate of the equipment at the same time, the scheduling objective function can be set as:

$\mathrm{<span\; class="notranslate">\; Obj</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; KT</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; 100</span>}<span\; class="notranslate">\; \%</span><span\; class="notranslate">\; )</span>$

Among them, c ₁ and c ₂ are constants. If the delivery date is more important, a larger c ₁ can be selected; if the equipment utilization rate is more important, a larger c ₂ can be selected.

2. Specify scheduling rules r _M,k and r _J,i for production equipment k and workpiece i respectively. A simplified processing method can also be adopted: specify a scheduling rule r _M for all equipment, and specify a scheduling rule r _J for all workpieces. A simpler approach is also possible: specify the same scheduling rule r for all equipment and workpieces. Commonly used scheduling rules are generally optional, such as: the longest remaining processing time priority, the shortest remaining processing time priority, the largest number of remaining processes priority, the least number of remaining processes priority, the earliest delivery date priority, the shortest slack time priority, and the highest processing efficiency priority wait;

3. Set the state and performance feedback coefficients in the iterative feedback scheduling rules (Formula 4a) and (Formula 4b) respectively:

p _k1 = p _k2 = q _k = 1, k = 1, 2, . . . , K

p _i =q _i1 =q _i2 =1, i=1, 2, . . . , N

If the scheduling objective function includes equipment utilization rate or equipment load unbalance rate, then increase q _k ; if the scheduling objective function includes the average workpiece delay time, or includes the maximum workpiece delay time, or includes the number of delayed workpieces, then increase q _i2 ; if the scheduling objective function includes the average processing time of the workpiece, or includes the average waiting time of the workpiece, or includes the maximum passing time of the workpiece, then increase q _i1 .

4. Set iteration control parameters and iteration initial value, namely:

Let the current number of iterations j=1, let the maximum number of iterations be N _max ;

Let S _opt be empty, and let Obj(S _opt ) be a sufficiently large number (1000 or 10000, etc.);

Let T ⁽⁰⁾ _k = 0, k = 1, 2, . . . , K;

Let t ⁽⁰⁾ _p,i =t ⁽⁰⁾ _w,i =t ⁽⁰⁾ _d,i =0,i=1,2,...,N

5. All equipment is set to idle state, and all workpieces are set to initial state.

6. According to the subroutine algorithm shown in Figure 3, allocate workpieces to be processed to idle equipment.

7. If there are unfinished workpieces, advance one unit of time and go to step 6.

8. End this iteration, and calculate the operating performance of this iteration system according to (Formula 2a) - (Formula 2h), calculate the scheduling objective function value Obj(S ^(j) ), j is the current iteration number, if Obj(S ^(j) )<Obj(S _opt ), then S _opt =S ^(j) .

9. Add 1 to the number of iterations. If j≤N _max , go to step 5 and enter the next iteration.

10. Stop the iteration, and output the optimal scheduling result S _opt and the corresponding scheduling objective function Obj(S _opt ) value.

Fig. 4 is a device composition and layout model (production system 1) of a production system implementing the scheduling algorithm of the present invention, and Fig. 5 is a schematic diagram of the communication connection of the system.

In the production system 1, the workpiece can be divided into rough R, semi-finished product, and finished product F according to the processing status of the workpiece. The blank is manually clamped on the standard pallet P at 4 loading and unloading stations, and enters the production system 1 to wait for processing; the finished product is at the 4 loading and unloading stations, and is manually removed and exits the production system 1. The loading and unloading station 4 is provided with a display device 2, which is used to receive and display the output command of the central controller 8, and prompt the operator to complete the type of operation currently: that is, remove the finished product on the pallet P at the loading and unloading station 4, and press the central controller The command of 8 requires that the blank of the specified type be clamped in the specified way on the pallet. There is also an input device 3 at the loading and unloading station 4, which is used for the operator to send information to the central controller 8 to notify the central controller 8 that the finished product on the pallet at the loading and unloading station has been removed and a new blank has been clamped as required. The blank can be loaded into the production system 1 by an automatic track trolley 9 for processing.

System buffer station 6 is used for temporary storage of workpieces. There are a certain number of physical stations 7a, 7b, 7c, . In the system buffer station controller 5, the relative position of the physical station relative to the absolute origin of the system buffer station 6 is recorded, and the state of the physical station (whether there is a pallet on it, whether there is a workpiece on the pallet, the number of the pallet and the number of the pallet on the pallet) is also recorded. workpiece number). According to the command of the buffer station controller 5, the buffer station can move the designated physical station to the pallet exchange station S02, S03 or S04 of the buffer station to prepare for the pallet pick/place operation. The buffer station controller 5 is connected to the central controller 8 through a local area network LAN, and can receive commands from the central controller 8, and can also feed back status information of the system buffer station 6 to the central controller 8.

The automatic track trolley 9 can move on the track 10, and the trolley 9 is equipped with an automatic tray picking/putting mechanism, which can complete the tray picking/putting operation. In production system 1, each pallet exchange station S01, S02, S03, ..., S15 is equipped with a position indicating element, and the position detection element installed on the automatic track trolley (9) can detect the position of each pallet exchange station. position, thereby realizing the precise positioning of the automatic track dolly (9). The car receives the command of the central controller 8 through the wireless communication device A and returns the status information of the car. According to the command issued by the central controller 8, the automatic track trolley 9 can move to the designated pallet exchange station S01, S02, S03, ..., S15 and complete the pallet picking/putting operation, so the blanks, semi-finished products and finished products pass through the automatic track The trolley 9 can move between the loading and unloading station 4, the system buffer station 6 and the production equipment M1, M2, M3, . . .

In the production system 1, there are several production equipments M1, M2, M3, . . . Different production equipments have different processing capabilities and can complete different processes. The equipment has three states, namely processing state, idle state, fault (including planned maintenance) state. Each device has a device controller C1, C2, C3, . . . responsible for monitoring and controlling the operation of the device. The equipment controllers C1, C2, C3, ... are connected to the central controller 8 through the local area network LAN, and can receive the commands and NC machining programs downloaded from the central controller 8, and complete the corresponding processing tasks, and can also upload the equipment status to the central controller 8.

The central controller 8 is composed of a computer, input and output circuits, communication devices and corresponding scheduling software, etc., and is responsible for monitoring the operating status of the production system and controlling the system.

The central controller 8 is connected to the loading and unloading station display device 2 and the loading and unloading station input device 3 respectively through the output and input interfaces, connected to the automatic track trolley 9 through the wireless communication device A, and controlled by the system buffer station controller 5 and production equipment through the local area network Devices C1, C2, C3, ... are connected.

The central controller 8 stores the delivery dates of all workpieces in the production task and its extended process sequence planning tree, and runs the scheduling program according to the scheduling method proposed by the present invention to realize the scheduling control of the production system 1, that is, control the iterative cycle, And judge whether all the iterations are over, if it is over, output the optimal scheduling result of the system. In each iterative cycle, the status of each workpiece in the system is recorded in real time, that is, whether the workpiece is in the processing state or to be processed, and the process of the workpiece that has just been processed (or is being processed) recently. In addition, the status of each device is recorded in real time, that is, whether the device is in a processing state, idle state or fault state, the processing procedure completed by each device and the start time and end time of the process. At each decision-making moment, according to the feedback scheduling rules, determine which process of which workpiece is assigned to which equipment for processing. After completing a task assignment, change the status of related workpieces and equipment in the system in real time, and report to the relevant equipment Issue a pallet (workpiece) exchange command.

The pallet (workpiece) exchange command format that central controller 8 sends is as follows:

LD Pxx Sxx; take pallet Pxx at pallet exchange station Sxx

UL Pxx Sxx; deposit pallet Pxx at pallet change station Sxx

After completing a task distribution, after determining that a certain process of a certain workpiece (such as workpiece i) is assigned to a certain equipment (such as equipment k) for processing, the central controller 8 will send a tray to the system buffer station controller 5 ( Workpiece i) command, buffer station controller 5 will move the pallet loaded with workpiece i in the buffer station to the pallet exchange station S02, S03 or S04 after receiving the command, and wait for the automatic track trolley 9 to take the pallet. At the same time, the central controller 8 will send an order to take pallets to the automatic track trolley 9, and after receiving the command, the trolley 9 will move to the designated buffer station pallet exchange station to take out the pallets. Thereafter, the central controller 8 will send a pallet storage command to the automatic track trolley 9. After receiving the command, the trolley 9 will store the pallet carrying the workpiece i to the device k for processing.

In addition, when a piece of equipment finishes processing a process, if the workpiece has unfinished processes, or all processes of the workpiece have been processed but there are pallets on the loading and unloading station, the central controller 8 will send a pallet pick-up call to the automatic track trolley 9. command, take out the pallet in the device and the workpiece clamped on the pallet, and send the pallet storage command to the automatic track trolley and the system buffer station, and send the pallet and the workpiece clamped on the pallet to the system buffer station 6 temporarily live. If all the processes of the workpiece have been processed and the current loading and unloading station is idle, the central controller 8 successively sends a pallet fetching command and a pallet storage command to the automatic track trolley 9 to take out the pallet in the device and the workpiece clamped on the pallet, Send it to the loading and unloading station of the system, and output display information to the loading and unloading station, informing the operator to unload the finished product on the pallet and clamp the new blank as required.

If there are finished products in the system buffer station 6, the central controller 8 will notify the automatic track trolley 9 to take out the finished products and send them to the loading and unloading station 4. Simultaneously, the central controller 8 outputs an order to the display device 2 of the loading and unloading station to notify the operator at the loading and unloading station 4 to take off the finished product on the pallet and clamp a new blank as required. If there is an idle pallet in the system buffer station 6, the central controller 8 notifies the automatic rail car 9 to take out the pallet and send it to the loading and unloading station 4. Simultaneously, the central controller 8 outputs commands to the loading and unloading station display device 2 to notify the operators at the loading and unloading station 4 to clamp new blanks as required.

After the blank is clamped at the 4 loading and unloading stations, the operator sends a signal through the loading and unloading station input device 3. After receiving the signal, the central controller 8 controls the automatic track trolley 9 to take the blank and send it to the system buffer 6 for processing.

A similar system of this production system is: on the basis of production system 1, add the system input buffer station and the system output buffer station. The input buffer station is used to store blanks, and the output buffer station is used to store finished products. Another similar system of this production system is: in front of each production equipment M1, M2, M3, ..., local buffer stations can be added as needed, similarly, local buffer stations can also be divided into input buffer stations and output buffer stations stand.

Claims (5)

1. A rule scheduling method with state performance feedback, comprising the steps in turn: (1) Formulate the scheduling objective function Obj according to the production task; (2) Designate scheduling rules r _{M, k} , r _J, i for production equipment k and workpiece i respectively, k=1, 2,..., K, i=1, 2,..., N; (3) Construct iterative feedback scheduling rules r _{cM, k} , r _{cJ, i} : r _{cM, k} = r _{M, k} +p _k1 R _k +p _k2 D _k +q _k T _k r _{cJ, i} = r _{J, i} + p _i t _{s, i} + q _i1 t _{p, i} /t _{w, i} -q _i2 t _{d, i} Among them, p _k1 and p _k2 are the state feedback coefficients of equipment k, q _k is the performance feedback coefficient of equipment k, p _i is the state feedback coefficient of workpiece i, q _i1 and q _i2 are the performance feedback coefficients of workpiece i; R _k is The demand of equipment k, D _k is the demand trend of equipment k, T _k is the processing time of equipment k; t _{s, i} is the average relaxation time of workpiece i, t _{p, i} is the processing time of workpiece i, t _{w , i} is the waiting time of job i, t _{d, i} is the delay time of job i; where, Equipment demand R _k (t) is calculated by the following formula: R _k (t)=|{J _i (t)}|, where J _i (t) represents the workpieces that can be processed on the production equipment at time t, {} represents the set, and || represents the basis of the set; The equipment demand trend D _k (t) is: ${\mathrm{D.}}_k\left(t\right)=\frac{d}{\mathrm{dt}}{f}_k\left(t\right),$ In the formula, F _k (t) is a continuous derivable function generated by fitting according to R _k (t); The average relaxation time t _{s of workpiece i, i} is: $t_{\mathrm{the\; s},i}=\frac{d_i-t-t_{r,i}}{{\mathrm{no}}_{r,i}},$ In the formula, t is the current moment, d _i is the planned delivery time of workpiece i, t _{r, i} is the expected value of the total processing time required for the unprocessed process of workpiece i, n _{r, i} is the number of unprocessed processes of workpiece i; (4) Set iteration control parameters and iteration initial value: Let the current number of iterations j=1, let the maximum number of iterations be N _max ; Let S _opt be empty, and let Obj(S _opt ) be a sufficiently large number; Let T ⁽⁰⁾ _k = 0, t ⁽⁰⁾ _{p, i} = t ⁽⁰⁾ _{w, i} = t ⁽⁰⁾ _{d, i} = 0; (5) All equipment is set to idle state, and all workpieces are set to initial state; (6) Assign workpieces to be processed to idle equipment according to the following steps: (6.1) If there is no idle equipment or no workpiece to be processed, go to step (7); (6.2) Utilize the following formula to calculate the priority of all idle devices; r ^(j) _{cM, k} = r _{M, k} +p _k1 R ^(j) _k +p _k2 D ^(j) _k +q _k T ^(j-1) _k (6.3) Complete each idle device in order of priority from high to low: (6.3.1) Determine the set L of workpieces to be processed that can be processed on it, if L is an empty set, then take the idle device with the second priority, and repeat this step; (6.3.2) Calculate the priority of workpieces to be processed in L according to the following formula: r ^(j) _{cJ, i} = r _{J, i} + p _i t ^(j) _{s, i} + q _i1 t ^(j-1) _{p, i} /t ^(j-1) _{w, i} -q _i2 t ^{(j -1)} _d,i (6.3.3) Take the workpiece with the highest priority and process it for the idle device; (6.3.4) judge whether there is also idle equipment, if so, enter step (6.3.1), otherwise, enter step (7); (7) If all workpieces are processed, turn to step (8); otherwise, advance one unit of time, turn to step (6); (8) At the end of this iteration, calculate the operating performance of this iteration system, and calculate the scheduling objective function value Obj(S ^(j) ), where j is the current iteration number, if Obj(S ^(j) )<Obj(S _opt ), Then S _opt = S ^(j) ; (9) Make j=j+1, if j≤N _max , go to step (5), otherwise go to step (10); (10) Stop the iteration and output the optimal scheduling result, which can be directly used for the scheduling of the same production task. 2. The scheduling method according to claim 1, characterized in that: step (1) formulates scheduling objective function Obj in the following manner: If the delivery time of the production task is tight or late delivery will be penalized, choose one of the following formulas as the scheduling objective function, namely: $\mathrm{Obj}=\min t_d=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}\left(\max \right\{t_{d,i},0\left\}\right),$ or Obj = mint _{d, max} = min(max{max{t _{d, i} }, 0}), or Obj＝minN _d ＝min|{i, t _{d, i} > 0}| If the delivery time of the production task is loose, choose one of the following formulas as the scheduling objective function, namely: $\text{Obj=maxU=max}\left(\frac{1}{\mathrm{KT}}\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{T}_k\right)\&times;100\%,$ or $\mathrm{Obj}=\min B=\min \left(\frac{1}{K}\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}\right|T_k-\frac{1}{K}\underset{j=1}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\left|\right)\&times;100\%,$ or $\mathrm{Obj}=\min t_p=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{p,i},$ or $\mathrm{Obj}=\min t_w=\min \frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{t}_{w,i},$ or Obj = mint _{t, max} = min(max {t _{p, i} + t _{w, i} }) When it is required to ensure the delivery date of the workpiece and improve the utilization rate of the equipment at the same time, the scheduling objective function is set as: $\mathrm{<span\; class="notranslate">\; Obj</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; KT</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; 100</span>}<span\; class="notranslate">\; \%</span><span\; class="notranslate">\; )</span>$ Where c ₁ and c ₂ are constants, t _d is the average workpiece delay time, K is the number of production equipment in the system, T is the scheduling cycle, and T _k is the processing time of equipment k. 3. The method according to claim 1 or 2, characterized in that: it also includes a method for determining key factors for measuring system performance, specifically including: the key factor for measuring equipment related performance is equipment processing time, and the key factor for measuring workpiece completion performance The key factors to measure the flow performance of workpieces are workpiece processing time and workpiece waiting time. 4. The method according to claim 1, characterized in that: using the extended process sequence planning tree to represent the process sequence planning flexibility of the workpiece and the processing path flexibility of the workpiece. 5. a dispatching system realizing the method according to claim 1, comprising central controller (8), loading and unloading station (4), system buffer station (6), automatic track dolly (9) and processing equipment, is characterized in that: The processing equipment is equipped with an equipment controller; The blank enters the system buffer station (6) through the loading and unloading station (4) to wait for processing, and the finished product exits the system through the loading and unloading station (4); The system buffer station (6) is used for temporarily storing workpieces, and it is provided with a buffer station controller (5); the system buffer station (6) is provided with physical stations, and each physical station is provided with a pallet, which is clamped to be Workpiece processing; the relative position of the physical station relative to the absolute origin of the system buffer station (6) is recorded in the buffer station controller (5), and the state of the physical station is also recorded; The automatic track car (9) moves between the loading and unloading station (4), the system buffer station (6) and the processing equipment, and an automatic pallet picking/putting mechanism is attached to it; The loading and unloading station (4), the system buffer station (6) and the processing equipment are all equipped with pallet exchange stations, and each pallet exchange station is equipped with a position indicating element, which is installed at the position on the automatic track trolley (9) The detection element is used to detect each pallet exchange station to realize the precise positioning of the automatic track trolley (9); The central controller (8) is respectively connected with the display device (2) of the loading and unloading station, the input device (3) of the loading and unloading station, the automatic track trolley (9), the system buffer station controller (5) and the equipment controller, and controls each component Work.

Images