A planning method for human-machine collaborative production tasks

Abstract

The present invention relates to the technical field of human-robot collaboration. The present invention provides a planning method for human-robot collaborative production tasks that provides workers with short breaks during the production process, and solves multi-objective optimization for the shortest production cycle and the minimum worker fatigue through a chemical reaction optimization method, so as to obtain the optimal balance between the worker's fatigue level and the production cycle. The task planning method provided by the present invention integrates the short rest time for workers, so that workers can get a short rest during the production process, reduce worker fatigue, and maximize the efficiency of the production line. The traditional model does not include a short rest time, and only provides workers with rest time when the worker's fatigue will exceed the fatigue constraint in the next production cycle. By comparison, the method provided by the present invention takes into account the worker's fatigue factor, and the solved work cycle is shorter, so the production efficiency is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of human-robot collaboration, and in particular to a method for planning a human-robot collaborative production task. Background Art

[0002] Driven by information technology, robotics will play an important role in intelligent manufacturing and flexible manufacturing. Assembly production, as an important part of the manufacturing process, accounts for 20-70% of the workload, 40-60% of the time, and 30-60% of the cost. The efficiency, cost and quality of assembly are directly related to the overall production cycle, cost and quality of the entire product. Traditional assembly is usually done manually, with poor efficiency and low reliability. The introduction of robots in assembly is more conducive to improving the automation level of the manufacturing industry.

[0003] As robots do not yet have the ability to complete complex production tasks, a human-machine collaborative system in which workers and robots form a hybrid system to share the workspace and jointly complete production tasks will be an important production model. Among them, workers undertake delicate and complex work tasks and spend more energy on supervision, decision-making and exception handling; while robots undertake highly repetitive and intensive work tasks.

[0004] As an important part of the hybrid system, people improve the system's ability to cope with complex tasks and abnormal situations on the one hand; on the other hand, they are also the shortcoming of the system's long-term stable operation. Workers cannot work continuously for a long time, and sufficient rest time must be arranged to ensure the physical and mental health of the workers, which limits the system's ability to operate efficiently and reliably for a long time. Therefore, although human-robot collaboration has strong capabilities, it also brings challenges to task planning between humans and robots. In task planning, workers' fatigue must be fully considered, and workers must be provided with appropriate rest and recovery time to ensure that collaborative production is carried out normally and orderly. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a planning method for human-robot collaborative production tasks that provides workers with a short break during the production process, and solves multi-objective optimization for the shortest production cycle and the least worker fatigue through a chemical reaction optimization method, and finds the optimal balance between the worker's fatigue level and the production cycle. The specific technical solution is as follows:

[0006] A planning method for a human-machine collaborative production task comprises the following steps:

[0007] S1: According to the attributes of the human-machine collaborative work unit, a production task planning model P(h,r) for the human-machine collaborative work unit is established, as follows:

[0008] P(h,r)=[A,h,r,T],T=[T H ,T R ];(1)

[0009] Among them, h is the number of workers in the human-machine collaborative work unit, r is the number of robots in the human-machine collaborative work unit, A is the production task that needs to be completed by human-machine collaboration, T is the total working time to complete a single production task A, and T H is the working time for a worker to complete a single production task A, T R The man-hours required for the robot to complete a single production task;

[0010] The production task planning model sets two fitness functions, Fit1 and Fit2, where Fit1 represents the sum of production time and rest time required for each production cycle; Fit2 represents the maximum fatigue value of a worker completing a production shift;

[0011] S2: Decompose production task A into n task sequences, namely:

[0012] A={S1,S2,…,S n}; (2)

[0013] Task sequences need to be executed in order, and the next task sequence can only be started after the previous task sequence is completed;

[0014] The i-th task sequence S i It can be decomposed into m i The subtasks are as follows:

[0015]

[0016] m i is the i-th task sequence S i The number of subtasks included, tk i,j is the i-th task sequence S i The jth subtask in i ; The i-th task sequence S i There is no order restriction on the subtasks in; workers complete the task sequence S i Each subtask in has the same fatigue coefficient and recovery coefficient;

[0017] S3: Worker H l Complete the i-th task sequence S i Working hours For workers l The sum of the time required to complete all subtasks in the task sequence is shown in the following formula:

[0018]

[0019] In the formula, For workers l Complete the i-th task sequence S i The working hours of the jth subtask;

[0020] Robot R k Complete the i-th task sequence S i Working hours For robot R k The sum of the time required to complete all subtasks in the task sequence is shown in the following formula:

[0021]

[0022] In the formula, For robot R k Complete the i-th task sequence S i The working hours of the jth subtask;

[0023] Complete the i-th task sequence S i Required working hours The maximum value of the working hours used by all workers and robots:

[0024]

[0025] S4: The fitness function Fit1 is obtained by the following formula:

[0026]

[0027] in, is the time required to complete the entire task sequence, that is, the time required to complete production task A; B is the time interval between cycles;

[0028] The fitness function Fit2 is obtained by the following method:

[0029] Assume that worker H l The completion time of the first subtask assigned is τ 1' , fatigue coefficient is λ 1' , then the fatigue value of the worker after completing the subtask is obtained by the following formula:

[0030]

[0031] In the formula, is the worker’s initial fatigue value; when the worker gets enough rest, at the beginning of each shift, the worker’s fatigue value is 0, that is, the worker’s fatigue value at the beginning of the first subtask in the first production cycle is 0; for other production cycles, the worker’s fatigue value at the beginning of the first subtask is the fatigue value after a short rest, see formula (10);

[0032] Assume that in each production cycle, worker H l Assigned to y subtasks, the fatigue value of the worker after completing the yth subtask is calculated by the following formula:

[0033]

[0034] In the formula, For workers l The fatigue value after completing the yth subtask; λ y' , τ y' are the fatigue coefficient corresponding to the y-th subtask and the working time required to complete it;

[0035] Worker H l After the break After rest, the remaining fatigue is as follows:

[0036]

[0037] Where μ is the worker's coefficient of recovery;

[0038] Worker H l Rest time for each production cycle It is determined by three parts of time, as shown in the following formula:

[0039]

[0040] Where, T be l For workers l The time from the moment the last subtask in a production cycle is completed to the moment the production cycle ends; T bs l For workers l The time from the moment a production cycle starts to the moment the first subtask starts;

[0041] Assume that the total number of production cycles completed in each shift is N, which is expressed by the following formula:

[0042]

[0043] Where W is the duration of each shift;

[0044]

[0045] S5: The chemical reaction algorithm is used to solve the minimum value of the two fitness functions Fit1 and Fit2, that is, the fatigue value of workers completing a production shift is minimized when the sum of the production time and rest time required for each production cycle is minimized, so as to obtain the optimal production task planning plan.

[0046] Preferably, the chemical reaction optimization algorithm comprises the following steps:

[0047] S51: Set the parameters of the chemical reaction algorithm, that is, set the initial kinetic energy loss rate L R , buffer energy E B 、Molecular kinetic energy K E , collision probability coll, decomposition threshold α, synthesis threshold β, population size G, number of iterations I, length of time coding b, weight coefficient ω1;

[0048] S52: Initialize the population of the chemical reaction algorithm, including:

[0049] Initialize the population Pop and encode each molecule. The encoding includes two parts: subtask encoding M and rest time encoding R, both of which are binary encodings.

[0050] Calculate the fitness function Fit1 corresponding to each molecule in the population;

[0051] Calculate the fitness function Fit2 corresponding to each molecule in the population;

[0052] Calculate the potential energy corresponding to each molecule in the population;

[0053] S53: Determine whether to perform a multi-molecule reaction or a single-molecule reaction; randomly select a random number. If the random number is greater than the collision probability coll or the population size is less than Then one molecule is randomly selected to perform a single-molecule reaction, and the process goes to step S54; otherwise, two molecules are randomly selected to perform a multi-molecule reaction, and the process goes to step S57;

[0054] S54: Determine whether to enter into a single molecule invalid collision reaction or a single molecule decomposition reaction; if the number of collisions of the molecule is greater than the decomposition threshold α, then a single molecule decomposition reaction is performed, and the process proceeds to step S55; otherwise, a single molecule invalid collision reaction is performed, and the process proceeds to step S56;

[0055] S55: Perform a unimolecular decomposition reaction and enter step S510 to update the state;

[0056] S56: Perform a single molecule invalid collision reaction and enter step S510 to update the state;

[0057] S57: Determine whether to enter into a multi-molecule intermolecular collision reaction or a multi-molecule molecular synthesis reaction; if the kinetic energy of both molecules is greater than the synthesis threshold β, then enter into a multi-molecule molecular synthesis reaction and enter step S58; otherwise, enter into a multi-molecule intermolecular collision reaction and enter step S59;

[0058] S58: Perform molecular synthesis reaction of multiple molecules and enter step S510 to update the state;

[0059] S59: perform a molecular collision reaction between multiple molecules and enter step S510 to update the state;

[0060] S510: Update the status, including updating the best molecule, and increase the number of iterations by 1;

[0061] S511: Determine whether the iteration is completed. If the number of iterations is less than I, enter step S53 to continue iterating; otherwise, exit the iteration and output the best molecule; obtain the optimal production task planning plan based on the subtask code M and rest time code R of the best molecule.

[0062] Preferably, the step of performing unimolecular decomposition reaction in step S55 comprises:

[0063] S551: Randomly generate a new task code M' and rest time code R';

[0064] S552: A new molecule is formed by combining the subtask code M and rest time code R of the original molecule with the newly generated M' and R', wherein molecule 1 is a combination of M and R', and molecule 2 is a combination of M' and R;

[0065] S553: ​​Calculate the fitness function Fit1 corresponding to the new molecule;

[0066] S554: Calculate the fitness function Fit2 corresponding to the new molecule;

[0067] S555: Calculate the potential energy PE of the new molecule;

[0068] S556: Update the kinetic energy KE of the new molecule;

[0069] S557: The number of collisions between the two new molecules is set to 0;

[0070] S558: The original molecule is deleted from the population;

[0071] S559: Add the two new molecules mentioned above to the population.

[0072] Preferably, the step of performing a single molecule invalid collision reaction in step S56 includes:

[0073] S561: Randomly select whether to encode the task or the rest time for invalid collision response;

[0074] S562: Calculate the fitness function Fit1 corresponding to the new molecule;

[0075] S563: Calculate the fitness function Fit2 corresponding to the new molecule;

[0076] S564: Calculate the potential energy PE of the new molecule;

[0077] S565: The number of collisions of the original molecule increases by 1;

[0078] S566: Calculate the kinetic energy KE of the new molecule;

[0079] S567: Update buffer energy E B ;

[0080] S568: The original molecule is deleted from the population;

[0081] S569: Add the two new molecules mentioned above to the population.

[0082] Preferably, the step of performing a molecular synthesis reaction of multiple molecules in step S58 includes:

[0083] S581: Randomly select a synthesis method from the following two methods: suppose the two molecules in step S57 are molecule 1 and molecule 2; the task code M1 of molecule 1 and the time code R2 of molecule 2 form a new molecule, and the task code M2 ​​of molecule 2 and the time code R1 of molecule 1 form a new molecule;

[0084] S582: Calculate the fitness function Fit1 corresponding to the new molecule;

[0085] S583: Calculate the fitness function Fit2 corresponding to the new molecule;

[0086] S584: Calculate the potential energy PE of the new molecule;

[0087] S585: Calculate the kinetic energy KE of the new molecule;

[0088] S586: Set the number of collisions of the new molecule;

[0089] S587: Add new molecules to the population;

[0090] S588: Remove the above two parent molecules from the population.

[0091] Preferably, the step of performing the intermolecular collision reaction of multiple molecules in step S59 includes:

[0092] S591: exchanging the task code columns of the two molecules in step S57;

[0093] S592: exchanging the rest time code sequences of the two molecules in step S57;

[0094] S593: The collision number of both molecules increases by 1;

[0095] S594: Calculate the fitness function Fit1 corresponding to the new molecule;

[0096] S595: Calculate the fitness function Fit2 corresponding to the new molecule;

[0097] S596: Calculate the potential energy PE of the new molecule;

[0098] S597: Calculate the kinetic energy KE of the new molecule;

[0099] S598: Two new molecules are added to the population;

[0100] S599: Remove the aforementioned two parent molecules from the population.

[0101] Preferably, the potential energy PE of the molecule is determined by fitness functions Fit1 and Fit2, as shown in the following formula:

[0102]

[0103] f1 - is the minimum value of all molecules Fit1 in the initial population; f1 + is the maximum value of all molecules Fit1 in the initial population; f2 - is the minimum value of Fit2 of all molecules in the initial population; f2 + is the maximum value of Fit2 of all molecules in the initial population.

[0104] Preferably, the kinetic energy KE of the new molecule in the unimolecular decomposition reaction in step S55 is specifically calculated as follows:

[0105] The temporary variable tmp1 is obtained by the following formula:

[0106] tmp1 = PE + KE - PE1' - PE2';

[0107] Where PE is the potential energy of the molecule before decomposition, KE is the kinetic energy of the molecule before decomposition, PE1' is the potential energy of the molecule after decomposition, and PE2' is the potential energy of the molecule after decomposition;

[0108] If tmp1+E B ≥0, then exit step S557, otherwise update the kinetic energy KE of the molecule and proceed to the next step S557, and update the kinetic energy KE of the new molecule specifically as follows:

[0109] Take a random number from 0 to 1 and assign it to q;

[0110] If tmp1 ≥ 0, the kinetic energy KE1' of the new molecule 1 is:

[0111] KE1' = tmp1*q;

[0112] The kinetic energy KE2' of the new molecule 2 is:

[0113] KE2' = tmp1*(1-q);

[0114] If tmp1<0, the kinetic energy KE1' and KE2' of the new molecule are obtained by the following formula:

[0115] KE1'=(tmp1+E B )*q;

[0116] KE2'=(tmp1+E B -KE1')*q.

[0117] Preferably, the kinetic energy KE of the new molecule in the single molecule invalid collision reaction in step S56 is specifically calculated as:

[0118] The temporary variable tmp2 is obtained by the following formula:

[0119] tmp2 = PE + KE - PE';

[0120] Where PE is the potential energy of the molecule before the collision, KE is the kinetic energy of the molecule before the collision, and PE' is the potential energy of the molecule after the collision;

[0121] If tmp2<0, then exit step S56;

[0122] If tmp2 ≥ 0, perform the following operations:

[0123] From L R From 1 to 1, take a random number and assign it to q, then:

[0124] The kinetic energy KE of the new molecule is tmp2*q;

[0125] Buffer energy E B Increase to tmp2*(1-q).

[0126] Preferably, the kinetic energy KE of the new molecule in the multi-molecule molecular synthesis reaction in step S58 is specifically calculated as:

[0127] The temporary variable tmp3 is obtained by the following formula:

[0128] tmp3 = PE1 + KE1 + PE2 + KE2 - PE';

[0129] Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, and PE' is the potential energy of the new molecule obtained after synthesis;

[0130] If tmp3>0, the kinetic energy of the new molecule KE=tmp3, otherwise tmp3≤0, the kinetic energy of the new molecule KE=0.

[0131] Preferably, the kinetic energy KE of the new molecule in the intermolecular collision reaction of multiple molecules in step S59 is specifically calculated as follows: the temporary variable tmp4 is obtained by the following formula:

[0132] tmp4=PE1+KE1+PE2+KE2-PE1'-PE2';

[0133] Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, PE1' is the potential energy of the new molecule 1; PE2' is the potential energy of the new molecule 2;

[0134] If tmp4<0, then exit the current step S59;

[0135] If tmp4 ≥ 0, perform the following operations:

[0136] Random numbers from 0 to 1 are assigned to q.

[0137] The kinetic energy of the new molecule 1 is tmp4*q

[0138] The kinetic energy of the new molecule 2 is assigned tmp4*(1-q).

[0139] The beneficial effects of the present invention are:

[0140] 1. The task planning method provided by the present invention integrates the short rest time of workers, so that during the production process, workers can get a short rest, reduce worker fatigue, and maximize the efficiency of the production line.

[0141] 2. The traditional model does not include short rest time, and only provides rest time for workers when their fatigue will exceed the fatigue constraint in the next production cycle. By comparison, the method provided by the present invention takes into account the fatigue factor of workers, and the working cycle is shorter, so the production efficiency is higher.

[0142] 3. The present invention adopts a chemical reaction algorithm in solving multi-objective task planning. The algorithm has strong robustness, fast convergence speed, and strong global search capability, which improves the accuracy of task planning solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0143] Figure 1 Flowchart for solving task planning based on chemical reaction optimization algorithm;

[0144] Figure 2 Example of a Gantt chart of workers’ work and rest time;

[0145] Figure 3 Example of binary coding of molecules;

[0146] Figure 4 Schematic diagram of single molecule invalid collision;

[0147] Figure 5 Schematic diagram of multi-molecule invalid collision;

[0148] Figure 6 A task planning Gantt chart obtained by the method of the present invention;

[0149] Figure 7 The task planning Gantt chart obtained by traditional methods;

[0150] Figure 8 A comparison chart of the production cycle of the task planning results obtained by the present invention and the traditional model.

[0151] Figure 2 In the , the molecular code is divided into two parts, task code and rest time code. The task code is divided into h+r groups, corresponding to h workers and r robots. In each group, there are S1 to S n sequence S i Contains m i The rest time code is b bits in total and is determined by initialization.

[0152] Figure 3 Middle, H 1 , H 2 , R 1 Represents the task planning Gantt charts for worker 1, worker 2, and robot 1 respectively. bs 1 is the rest time of worker 1 from the beginning of the cycle to the start of the first subtask in a cycle, T be 1 is the rest time of worker 1 from the end of the last subtask to the start of rest interval B in a cycle. w h1 is the time between the start of the first subtask and the end of the last subtask for worker 1 in a cycle. B is the rest interval after work. T bs 2 is the rest time of worker 2 from the beginning of the cycle to the start of the first subtask in a cycle, T be 2 is the rest time of worker 2 from the end of the last subtask to the start of rest interval B in a cycle. w h2 The time between the start of the first subtask and the end of the last subtask for worker 2 in a cycle.

[0153] Figure 4 (a) is a single molecule invalid collision coded for the task. The sequence S is randomly selected i The jth subtask in is taken as the modification object. A new set of subtask codes is randomly generated, the number of which is h+r, corresponding to the number of task code groups. The selected modification object is replaced with the newly generated subtask code to generate a new molecule. Figure 4 (b) is a single molecule invalid collision for rest time coding, and B in the coding is randomly selected. i Modify the B i To invert the original code, if it is 1, the new code is 0; if it is 0, the new code is 1.

[0154] Figure 5 (a) is a multi-molecule invalid collision for the task code. Randomly select the sequence S in the task code i The encoding of the j-th subtask in is swapped, and the encoding of the corresponding subtask between molecule 1 and molecule 2 is swapped to generate two new molecules. Figure 5 (b) Multi-molecule invalid collisions coded for rest time. Randomly select B in the code i The positions are swapped to generate new molecules.

[0155] Figure 6 The task planning Gantt chart obtained by the method of the present invention is shown in the figure for the case where the number of workers and robots is 2 respectively, and the time coding part of the obtained result is 0.

[0156] Figure 7 The task planning Gantt chart obtained by traditional methods. The chart is divided into two parts: work and rest. The results in the figure are for the case where the number of workers and robots are 2 respectively.

[0157] Figure 8 This is a comparison chart of the production cycle of the task planning results obtained by the present invention and the traditional model. The horizontal axis in the figure corresponds to the configuration of the number of robots and workers, and the vertical axis is the percentage of the production cycle improvement of the task planning results obtained by the present invention relative to the traditional model. DETAILED DESCRIPTION

[0158] In order to better understand the present invention, the present invention is further described below in conjunction with the accompanying drawings and specific embodiments:

[0159] The present invention is based on a human-machine collaborative task planning model that integrates short rest times. The production tasks are decomposed into subtasks and assigned to workers and robots to solve the multi-objective task planning results with the shortest task cycle time and the least worker fatigue. First, a task planning model for human-machine collaboration is defined; shortest working hours and minimum worker fatigue are defined as objective functions; the parameters of the chemical reaction algorithm are set and the population is initialized; then the objective function value of each molecule in the population is calculated; the chemical reaction optimization solution is performed on the molecules in the population; and finally, it is determined whether the termination condition is met. If so, the result is output, otherwise, the above steps are repeated to continue solving. A planning method for human-machine collaborative production tasks, comprising the following steps:

[0160] S1: According to the attributes of the human-machine collaborative work unit, a production task planning model P(h,r) for the human-machine collaborative work unit is established, as follows:

[0161] P(h,r)=[A,h,r,T],T=[T H ,T R ];(1)

[0162] Among them, h is the number of workers in the human-machine collaborative work unit, r is the number of robots in the human-machine collaborative work unit, A is the production task that needs to be completed by human-machine collaboration, T is the total working time to complete a single production task A, and T H is the working time for a worker to complete a single production task A, T R The man-hours required for the robot to complete a single production task;

[0163] The production task planning model sets two fitness functions, Fit1 and Fit2, where Fit1 represents the sum of production time and rest time required for each production cycle; Fit2 represents the maximum fatigue value of a worker completing a production shift.

[0164] S2: Decompose production task A into n task sequences, namely:

[0165] A={S1,S2,…,S n}; (2)

[0166] Task sequences need to be executed in order, and the next task sequence can only be started after the previous task sequence is completed;

[0167] The i-th task sequence S i It can be decomposed into m i The subtasks are as follows:

[0168]

[0169] m i is the i-th task sequence S i The number of subtasks included, tk i,j is the i-th task sequence S i The jth subtask in i ; The i-th task sequence S i There is no order restriction on the subtasks in; workers complete the task sequence S i Each subtask in has the same fatigue coefficient and recovery coefficient.

[0170] The decomposed task sequence and subtasks in this embodiment are shown in Table 1 below:

[0171] Table 1 Task decomposition and completion time information table

[0172]

[0173]

[0174] The task information in Table 1 includes: the number of sequences, the number of subtasks in each sequence, the type of each subtask, and the time required for each subtask to be completed by the robot and the worker. The time required to complete a single production task can be obtained according to the method in the following step S3.

[0175] S3: Worker H l Complete the i-th task sequence S i Working hours For workers l The sum of the time required to complete all subtasks in the task sequence is shown in the following formula:

[0176]

[0177] In the formula, For workers l Complete the i-th task sequence S i The working hours of the jth subtask; when tk i,j Assigned to worker H l ,but Equal to the t corresponding to the task h , the specific value corresponds to the worker's working hours t in Table 1 h ; When tk i,j Not assigned to worker H l but is equal to 0.

[0178] Robot R k Complete the i-th task sequence S i Working hours For robot R k The sum of the time required to complete all subtasks in the task sequence is shown in the following formula:

[0179]

[0180] In the formula, For robot R k Complete the i-th task sequence S i The working hours of the jth subtask; when tk i,j Assigned to robot R k ,but Equal to the t corresponding to the task r , the specific values ​​correspond to the robot working hours t in Table 1 r ; When tk i,j Not assigned to robot R k but is equal to 0.

[0181] Complete the i-th task sequence S i Required working hours The maximum value of the working hours used by all workers and robots:

[0182]

[0183] In the formula, the max() function is used to find the maximum value of all the values ​​in the brackets.

[0184] S4: The multi-objective task planning designed by the present invention includes two fitness functions, namely time Fit1 and worker fatigue value Fit2. Among them, Fit1 represents the sum of production time and rest time required for each production cycle; Fit2 represents the maximum fatigue value of a worker completing a production shift.

[0185] The fitness function Fit1 is obtained by the following formula:

[0186]

[0187] in, is the time required to complete the entire task sequence, that is, the time required to complete production task A; B is the time interval between cycles; Figure 2 shown.

[0188] The fitness function Fit2 is obtained by the following method:

[0189] Workers will become fatigued when completing their work. The fatigue coefficient and recovery coefficient of workers completing each subtask are shown in Table 2. The specific calculation is as follows:

[0190] Table 2 Fatigue coefficient and recovery coefficient corresponding to subtasks

[0191] Task number λ μ Task number λ μ 1 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 8 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 2 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 9 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 3 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 10 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 4 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 11 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 5 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 12 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 6 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 13 <![CDATA[1.1×10 -4 ]]> <![CDATA[1.6×10 -4 ]]> 7 <![CDATA[1.6×10 -4 ]]> <![CDATA[1.6×10 -4 ]]>

[0192] Assume that worker H l The completion time of the first subtask assigned is τ 1 ', fatigue coefficient is λ 1 ', in order to distinguish the "first assigned subtask" from the "first task" mentioned above, the superscript 1' is used here to distinguish them.

[0193] The worker's fatigue after completing the subtask is calculated by the following formula:

[0194]

[0195] In the formula, is the worker’s initial fatigue value; when the worker gets enough rest, at the beginning of each shift, the worker’s fatigue value is 0, that is, the worker’s fatigue value at the beginning of the first subtask in the first production cycle is 0; for other production cycles, the worker’s fatigue value at the beginning of the first subtask is the fatigue value after a short rest, see formula (10);

[0196] Assume that in each production cycle, worker H l Assigned to y subtasks, the worker's fatigue after completing the yth subtask is calculated by the following formula:

[0197]

[0198] In the formula, For workers l The fatigue value after completing the yth subtask; λ y '、 τ y 'respectively the fatigue coefficient corresponding to the y-th subtask and the working time required to complete it;

[0199] Worker H l After the break After resting, the remaining fatigue value is as follows:

[0200]

[0201] Where μ is the worker's coefficient of recovery;

[0202] Worker H l Rest time for each production cycle It is determined by three parts of time, as shown in the following formula:

[0203]

[0204] Where, T be l For workers l The time from the moment the last subtask in a production cycle is completed to the moment the production cycle ends; T bs l For workers l The time from the start of a production cycle to the start of the first subtask; T be l , T bs l Depend on Figure 3 Label.

[0205] Assume that the total number of production cycles completed in each shift is N, which is expressed by the following formula:

[0206]

[0207] Where W is the duration of each shift;

[0208]

[0209] In the formula, the max() function is used to find the maximum value of all items in the brackets; For workers l The fatigue value is obtained by iteratively calculating equations (8) and (9), where the number of iterations is the total number of cycles N completed by each shift.

[0210] S5: The chemical reaction algorithm is used to solve the minimum value of the two fitness functions Fit1 and Fit2, that is, the fatigue value of workers completing a production shift is minimized when the sum of the production time and rest time required for each production cycle is minimized, so as to obtain the optimal production task planning plan.

[0211] The chemical reaction optimization algorithm consists of the following steps:

[0212] S51: Set the parameters of the chemical reaction algorithm, that is, set the initial kinetic energy loss rate L R =0.4, buffer zone energy E B =0, molecular kinetic energy K E =1, collision probability coll=0.3, decomposition threshold α=40, synthesis threshold β=0.25, population size G=1000, iteration number I=10000, time coding length b=5, weight coefficient ω1=0.5;

[0213] S52: Initialize the population of the chemical reaction algorithm, including:

[0214] S521: Initialize the population Pop and encode each molecule. The encoding includes two parts: subtask encoding M and rest time encoding R, both of which are binary encodings; Figure 2 As shown;

[0215] S522: Calculate the fitness function Fit1 corresponding to each molecule in the population; obtained by the above formula (7);

[0216] S523: Calculate the fitness function Fit2 corresponding to each molecule in the population; obtained by the above formula (13);

[0217] S524: Calculate the potential energy corresponding to each molecule in the population;

[0218] The potential energy PE of a molecule is determined by the fitness functions Fit1 and Fit2, as shown in the following formula:

[0219]

[0220] f1- is the minimum value of all molecules Fit1 in the initial population; f1 + is the maximum value of all molecules Fit1 in the initial population; f2 - is the minimum value of Fit2 of all molecules in the initial population; f2 + is the maximum value of all molecules Fit2 in the initial population. Use the minimum value of all molecular fitness functions Fit1 to give f1 - Assign value, use the maximum value of all molecular fitness functions Fit1 to f1 + Assignment; use the minimum value of all molecular fitness functions Fit2 to f2 - Assign value, use the maximum value of all molecular fitness functions Fit2 to f2 + Assignment;

[0221] S53: Determine whether to perform a multi-molecule reaction or a single-molecule reaction; randomly select a random number. If the random number is greater than the collision probability coll or the population size is less than Then one molecule is randomly selected to perform a single-molecule reaction, and the process goes to step S54; otherwise, two molecules are randomly selected to perform a multi-molecule reaction, and the process goes to step S57;

[0222] S54: Determine whether to enter into a single molecule invalid collision reaction or a single molecule decomposition reaction; if the number of collisions of the molecule is greater than the decomposition threshold α, then a single molecule decomposition reaction is performed, and the process proceeds to step S55; otherwise, a single molecule invalid collision reaction is performed, and the process proceeds to step S56;

[0223] S55: Perform a unimolecular decomposition reaction and enter step S510 to update the state; the steps of performing a unimolecular decomposition reaction in step S55 include:

[0224] S551: Randomly generate a new task code M' and rest time code R';

[0225] S552: A new molecule is formed by combining the subtask code M and rest time code R of the original molecule with the newly generated M' and R', wherein molecule 1 is a combination of M and R', and molecule 2 is a combination of M' and R;

[0226] S553: ​​Calculate the fitness function Fit1 corresponding to the new molecule; as in the aforementioned step S522;

[0227] S554: Calculate the fitness function Fit2 corresponding to the new molecule; as in the aforementioned step S523;

[0228] S555: Calculate the potential energy PE of the new molecule; as in the aforementioned step S524;

[0229] S556: Update the kinetic energy KE of the new molecule; the specific calculation is as follows:

[0230] The temporary variable tmp1 is obtained by the following formula:

[0231] tmp1 = PE + KE - PE1' - PE2';

[0232] Where PE is the potential energy of the molecule before decomposition, KE is the kinetic energy of the molecule before decomposition, PE1' is the potential energy of the molecule after decomposition, and PE2' is the potential energy of the molecule after decomposition;

[0233] If tmp1+E B ≥0, then exit step S557, otherwise update the kinetic energy KE of the molecule and proceed to the next step S557, and update the kinetic energy KE of the new molecule specifically as follows:

[0234] Take a random number from 0 to 1 and assign it to q;

[0235] If tmp1 ≥ 0, the kinetic energy KE1' of the new molecule 1 is:

[0236] KE1' = tmp1*q;

[0237] The kinetic energy KE2' of the new molecule 2 is:

[0238] KE2' = tmp1*(1-q);

[0239] If tmp1<0, the kinetic energy KE1' and KE2' of the new molecule are obtained by the following formula:

[0240] KE1'=(tmp1+E B )*q;

[0241] KE2'=(tmp1+E B -KE1')*q.

[0242] S557: The number of collisions between the two new molecules is set to 0;

[0243] S558: The original molecule is deleted from the population;

[0244] S559: Add the two new molecules mentioned above to the population.

[0245] S56: Perform a single molecule invalid collision reaction and enter step S510 to update the state; the steps of performing a single molecule invalid collision reaction in step S56 include:

[0246] S561: Randomly select the task code or the rest time code to perform an invalid collision response; take a random number, if the random number is greater than 0.5, perform an invalid collision response on the task code, and randomly select the sequence S i The j-th subtask of is re-encoded as Figure 4.a; otherwise, perform an invalid collision reaction on the rest time code, randomly select a time code, and invert it. If the original value is 1, change it to 0; if the original value is 0, change it to 1, such as Figure 4 .b shown;

[0247] S562: Calculate the fitness function Fit1 corresponding to the new molecule; as in the aforementioned step S522;

[0248] S563: Calculate the fitness function Fit2 corresponding to the new molecule; as in the aforementioned step S523;

[0249] S564: Calculate the potential energy PE of the new molecule; as in the aforementioned step S524;

[0250] S565: The number of collisions of the original molecule increases by 1;

[0251] S566: Calculate the kinetic energy KE of the new molecule; the specific calculation is:

[0252] The temporary variable tmp2 is obtained by the following formula:

[0253] tmp2 = PE + KE - PE';

[0254] Where PE is the potential energy of the molecule before the collision, KE is the kinetic energy of the molecule before the collision, and PE' is the potential energy of the molecule after the collision;

[0255] If tmp2<0, then exit step S56;

[0256] If tmp2 ≥ 0, perform the following operations:

[0257] From L R From 1 to 1, take a random number and assign it to q, then:

[0258] The kinetic energy KE of the new molecule is tmp2*q;

[0259] Buffer energy E B Increase to tmp2*(1-q).

[0260] S567: Update buffer energy E B ;

[0261] S568: The original molecule is deleted from the population;

[0262] S569: Add the two new molecules mentioned above to the population.

[0263] S57: Determine whether to enter into a multi-molecule intermolecular collision reaction or a multi-molecule molecular synthesis reaction; if the kinetic energy of both molecules is greater than the synthesis threshold β, then enter into a multi-molecule molecular synthesis reaction and enter step S58; otherwise, enter into a multi-molecule intermolecular collision reaction and enter step S59;

[0264] S58: Performing a molecular synthesis reaction of multiple molecules, and entering step S510 to update the state; the steps of performing a molecular synthesis reaction of multiple molecules in step S58 include:

[0265] S581: Randomly select a synthesis method from the following two methods: suppose the two molecules in step S57 are molecule 1 and molecule 2 respectively; the task code M1 of molecule 1 and the time code R2 of molecule 2 form a new molecule, and the task code M2 ​​of molecule 2 and the time code R1 of molecule 1 form a new molecule; take a random number, if the random number is less than 0.5, the task code M1 of molecule 1 and the rest time code R2 of molecule 2 in step S57 form a new molecule; otherwise, the task code M2 ​​of molecule 2 and the rest time code R1 of molecule 1 form a new molecule;

[0266] S582: Calculate the fitness function Fit1 corresponding to the new molecule; as in the aforementioned step S522;

[0267] S583: Calculate the fitness function Fit2 corresponding to the new molecule; as in the aforementioned step S523;

[0268] S584: Calculate the potential energy PE of the new molecule; as in the aforementioned step S524;

[0269] S585: Calculate the kinetic energy KE of the new molecule; the specific calculation is:

[0270] The temporary variable tmp3 is obtained by the following formula:

[0271] tmp3 = PE1 + KE1 + PE2 + KE2 - PE';

[0272] Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, and PE' is the potential energy of the new molecule obtained after synthesis;

[0273] If tmp3>0, the kinetic energy of the new molecule KE=tmp3, otherwise tmp3≤0, the kinetic energy of the new molecule KE=0.

[0274] S586: Set the number of collisions of the new molecule; the number of collisions of the new molecule is equal to the average number of collisions of the original two molecules rounded off;

[0275] S587: Add new molecules to the population;

[0276] S588: Remove the above two parent molecules from the population.

[0277] S59: Performing a molecular collision reaction among multiple molecules, and entering step S510 to update the state; the steps of performing a molecular collision reaction among multiple molecules in step S59 include:

[0278] S591: Exchange the task code sequences of the two molecules described in step S57; Figure 5 .a shown;

[0279] S592: Exchange the rest time code sequences of the two molecules in step S57; Figure 5 As shown in .b, two new molecules are obtained;

[0280] S593: The collision numbers of the two molecules are both increased by 1; specifically, the new molecule 1 inherits the collision number of the parent molecule 1 and increases by 1, and the new molecule 2 inherits the collision number of the parent molecule 2 and increases by 1;

[0281] S594: Calculate the fitness function Fit1 corresponding to the new molecule; as in the aforementioned step S522;

[0282] S595: Calculate the fitness function Fit2 corresponding to the new molecule; as in the aforementioned step S523;

[0283] S596: Calculate the potential energy PE of the new molecule; as in the aforementioned step S524;

[0284] S597: Calculate the kinetic energy KE of the new molecule; the specific calculation is:

[0285] The temporary variable tmp4 is obtained by the following formula:

[0286] tmp4=PE1+KE1+PE2+KE2-PE1'-PE2';

[0287] Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, PE1' is the potential energy of the new molecule 1; PE2' is the potential energy of the new molecule 2;

[0288] If tmp4<0, then exit the current step S59;

[0289] If tmp4 ≥ 0, perform the following operations:

[0290] Random numbers from 0 to 1 are assigned to q.

[0291] The kinetic energy of the new molecule 1 is tmp4*q

[0292] The kinetic energy of the new molecule 2 is assigned tmp4*(1-q).

[0293] S598: Two new molecules are added to the population;

[0294] S599: Remove the aforementioned two parent molecules from the population.

[0295] S510: Update the status, including updating the best molecule, increasing the number of iterations by 1; the molecule with the smallest potential energy value is the best molecule; save the potential energy value of the best molecule and the corresponding task code and rest time code;

[0296] S511: Determine whether the iteration is completed. If the number of iterations is less than I, enter step S53 to continue iterating; otherwise, exit the iteration and output the best molecule; obtain the optimal production task planning plan based on the task code M and rest time code R of the best molecule.

[0297] In the traditional task planning model, the worker's fatigue factor is not considered during task planning, and only the shortest production cycle is considered. After obtaining the planning results, the worker's fatigue is predicted. When the worker's fatigue will exceed the limit in the next production cycle, the worker is provided with rest time to reduce fatigue, ensuring that the worker's fatigue will not exceed the limit in each production cycle.

[0298] The production cycle corresponding to the traditional task planning model can be expressed as:

[0299] Where, T W ' is the shortest production cycle time obtained without fatigue limitation, T B ' is the recovery time required for workers after completing a production cycle, The planning results include Figure 3 Similar worker breaks.

[0300] Figure 6 A task planning Gantt chart obtained by using the method of the present invention is given; Figure 7 The task planning Gantt chart obtained by traditional method is given; Figure 6 and Figure 7 The number of workers and robots in the process is 2 respectively. However, the production cycle time obtained by the method of the present invention is 164 seconds, while the production cycle time obtained by the traditional model method is 587 seconds.

[0301] It can be seen that the production task planning method provided in this application significantly shortens the production cycle compared to the traditional model, and workers can have a short break during the cycle, specifically, robots 1 and 2 can rest when executing task sequences S1, S4, S5 and robot 1 can rest when executing task sequence S6.

[0302] The comparison between the production cycle of the task planning results obtained by the present invention and the production cycle of the traditional model is shown in Table 3. The improvement ratio is as follows: Figure 8 As shown, it can be seen that the present invention is more optimized than the traditional production task planning model. Among them, the improvement ratio calculation method is: (the cycle obtained by the traditional method-the cycle obtained by the present invention) / the cycle obtained by the traditional method.

[0303] Table 3 Comparison of the production cycle of the task planning results obtained by the present invention and the production cycle of the traditional model

[0304] Production Model Cycles obtained by traditional method (s) The patent cycle (s) Increased ratio (%) P(1，1) 359 236 34.3 P(1，2) 580 196 66.2 P(2，1) 450 204 54.7 P(2，2) 587 164 72.1

[0305] The present invention is not limited to the specific implementation methods described above. The above is only a preferred implementation case of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A planning method for human-machine collaborative production tasks, characterized in that: The following steps are involved: S1: According to the attributes of the human-machine collaborative work unit, a production task planning model P(h,r) for the human-machine collaborative work unit is established, as follows: P(h,r)=[A,h,r,T],T=[T H ,T R ];(1) Among them, h is the number of workers in the human-machine collaborative work unit, r is the number of robots in the human-machine collaborative work unit, A is the production task that needs to be completed by human-machine collaboration, T is the total working time to complete a single production task A, and T H is the working time for a worker to complete a single production task A, T R The man-hours required for the robot to complete a single production task; The production task planning model sets two fitness functions, Fit1 and Fit2, where Fit1 represents the sum of production time and rest time required for each production cycle; Fit2 represents the maximum fatigue value of a worker completing a production shift; S2: Decompose production task A into n task sequences, namely: A={S1,S2,…,S n };(2) Task sequences need to be executed in order, and the next task sequence can only be started after the previous task sequence is completed; The i-th task sequence S i It can be decomposed into m i The subtasks are as follows: m i is the i-th task sequence S i The number of subtasks included, tk i,j is the i-th task sequence S i The jth subtask in i ; The i-th task sequence S i There is no order restriction on the subtasks in; workers complete the task sequence S i Each subtask in has the same fatigue coefficient and recovery coefficient; S3: Worker H l Complete the i-th task sequence S i Working hours For workers l The sum of the time required to complete all subtasks in the task sequence is shown in the following formula: In the formula, For workers l Complete the i-th task sequence S i The working hours of the jth subtask; Robot R k Complete the i-th task sequence S i Working hours For robot R k The sum of the time required to complete all subtasks in the task sequence is shown in the following formula: In the formula, For robot R k Complete the i-th task sequence S i The working hours of the jth subtask; Complete the i-th task sequence S i Required working hours The maximum value of the working hours used by all workers and robots: S4: The fitness function Fit1 is obtained by the following formula: in, is the time required to complete the entire task sequence, that is, the time required to complete production task A; B is the time interval between cycles; The fitness function Fit2 is obtained by the following method: Assume that worker H l The completion time of the first subtask assigned is τ 1 ', fatigue coefficient is λ 1 ', then the fatigue value of the worker after completing the subtask is obtained by the following formula: In the formula, is the worker’s initial fatigue value; when the worker gets enough rest, at the beginning of each shift, the worker’s fatigue value is 0, that is, the worker’s fatigue value at the beginning of the first subtask in the first production cycle is 0; for other production cycles, the worker’s fatigue value at the beginning of the first subtask is the fatigue value after a short rest, see formula (10); Assume that in each production cycle, worker H l Assigned to y subtasks, the fatigue value of the worker after completing the yth subtask is calculated by the following formula: In the formula, For workers l The fatigue value after completing the yth subtask; λ y '、 τ y 'respectively the fatigue coefficient corresponding to the y-th subtask and the working time required to complete it; Worker H l After the break After rest, the remaining fatigue is as follows: Where μ is the worker's coefficient of recovery; Worker H l Rest time for each production cycle It is determined by three parts of time, as shown in the following formula: Where, T be l For workers l The time from the moment the last subtask in a production cycle is completed to the moment the production cycle ends; T bs l For workers l The time from the moment a production cycle starts to the moment the first subtask starts; Assume that the total number of production cycles completed in each shift is N, which is expressed by the following formula: Where W is the duration of each shift; S5: The chemical reaction algorithm is used to solve the minimum value of the two fitness functions Fit1 and Fit2, that is, the fatigue value of workers completing a production shift is minimized when the sum of the production time and rest time required in each production cycle is minimized, so as to obtain the optimal production task planning scheme; The chemical reaction algorithm comprises the following steps: S51: Set the parameters of the chemical reaction algorithm, that is, set the initial kinetic energy loss rate L R , buffer energy E B 、Molecular kinetic energy K E , collision probability coll, decomposition threshold α, synthesis threshold β, population size G, number of iterations I, length of time coding b, weight coefficient ω1; S52: Initialize the population of the chemical reaction algorithm, including: Initialize the population Pop and encode each molecule. The encoding includes two parts: subtask encoding M and rest time encoding R, both of which are binary encodings. Calculate the fitness function Fit1 corresponding to each molecule in the population; Calculate the fitness function Fit2 corresponding to each molecule in the population; Calculate the potential energy corresponding to each molecule in the population; S53: Determine whether to perform a multi-molecule reaction or a single-molecule reaction; randomly select a random number. If the random number is greater than the collision probability coll or the population size is less than Then one molecule is randomly selected to perform a single-molecule reaction, and the process goes to step S54; otherwise, two molecules are randomly selected to perform a multi-molecule reaction, and the process goes to step S57; S54: Determine whether to enter into a single molecule invalid collision reaction or a single molecule decomposition reaction; if the number of collisions of the molecule is greater than the decomposition threshold α, then a single molecule decomposition reaction is performed, and the process proceeds to step S55; otherwise, a single molecule invalid collision reaction is performed, and the process proceeds to step S56; S55: Perform a unimolecular decomposition reaction and enter step S510 to update the state; S56: Perform a single molecule invalid collision reaction and enter step S510 to update the state; S57: Determine whether to enter into a multi-molecule intermolecular collision reaction or a multi-molecule molecular synthesis reaction; if the kinetic energy of both molecules is greater than the synthesis threshold β, then enter into a multi-molecule molecular synthesis reaction and enter step S58; otherwise, enter into a multi-molecule intermolecular collision reaction and enter step S59; S58: Perform molecular synthesis reaction of multiple molecules and enter step S510 to update the state; S59: perform a molecular collision reaction between multiple molecules and enter step S510 to update the state; S510: Update the status, including updating the best molecule, and increase the number of iterations by 1; S511: Determine whether the iteration is completed. If the number of iterations is less than I, enter step S53 to continue iterating; otherwise, exit the iteration and output the best molecule; obtain the optimal production task planning plan based on the subtask code M and rest time code R of the best molecule.

2. The method for planning a human-machine collaborative production task according to claim 1, characterized in that: The step of performing the unimolecular decomposition reaction in step S55 includes: S551: Randomly generate a new task code M' and rest time code R'; S552: A new molecule is formed by combining the subtask code M and rest time code R of the original molecule with the newly generated M' and R', wherein molecule 1 is a combination of M and R', and molecule 2 is a combination of M' and R; S553: ​​Calculate the fitness function Fit1 corresponding to the new molecule; S554: Calculate the fitness function Fit2 corresponding to the new molecule; S555: Calculate the potential energy PE of the new molecule; S556: Update the kinetic energy KE of the new molecule; S557: The number of collisions between the two new molecules is set to 0; S558: The original molecule is deleted from the population; S559: Add the two new molecules mentioned above to the population.

3. The method for planning a human-machine collaborative production task according to claim 1, characterized in that: The step of performing the single molecule invalid collision reaction in step S56 includes: S561: Randomly select whether to encode the task or the rest time for invalid collision response; S562: Calculate the fitness function Fit1 corresponding to the new molecule; S563: Calculate the fitness function Fit2 corresponding to the new molecule; S564: Calculate the potential energy PE of the new molecule; S565: The number of collisions of the original molecule increases by 1; S566: Calculate the kinetic energy KE of the new molecule; S567: Update buffer energy E B ; S568: The original molecule is deleted from the population; S569: Add the two new molecules mentioned above to the population.

4. The method for planning a human-machine collaborative production task according to claim 1, characterized in that: The steps of performing the molecular synthesis reaction of multiple molecules in step S58 include: S581: Randomly select a synthesis method from the following two methods: suppose the two molecules in step S57 are molecule 1 and molecule 2; the task code M1 of molecule 1 and the time code R2 of molecule 2 form a new molecule, and the task code M2 ​​of molecule 2 and the time code R1 of molecule 1 form a new molecule; S582: Calculate the fitness function Fit1 corresponding to the new molecule; S583: Calculate the fitness function Fit2 corresponding to the new molecule; S584: Calculate the potential energy PE of the new molecule; S585: Calculate the kinetic energy KE of the new molecule; S586: Set the number of collisions of the new molecule; S587: Add new molecules to the population; S588: Remove the above two parent molecules from the population.

5. The method for planning a human-machine collaborative production task according to claim 1, characterized in that: The step of performing the intermolecular collision reaction of multiple molecules in step S59 includes: S591: exchanging the task code columns of the two molecules in step S57; S592: exchanging the rest time code sequences of the two molecules in step S57; S593: The collision number of both molecules increases by 1; S594: Calculate the fitness function Fit1 corresponding to the new molecule; S595: Calculate the fitness function Fit2 corresponding to the new molecule; S596: Calculate the potential energy PE of the new molecule; S597: Calculate the kinetic energy KE of the new molecule; S598: Two new molecules are added to the population; S599: Remove the aforementioned two parent molecules from the population.

6. A method for planning a human-machine collaborative production task according to any one of claims 1 to 5, characterized in that: The potential energy PE of the molecule is determined by the fitness functions Fit1 and Fit2, as shown in the following formula: f1 - is the minimum value of all molecules Fit1 in the initial population; f1 + is the maximum value of all molecules Fit1 in the initial population; f2 - is the minimum value of Fit2 of all molecules in the initial population; f2 + is the maximum value of Fit2 of all molecules in the initial population.

7. The method for planning a human-machine collaborative production task according to claim 2, characterized in that: The kinetic energy KE of the new molecule in the unimolecular decomposition reaction in step S55 is specifically calculated as follows: The temporary variable tmp1 is obtained by the following formula: tmp1 = PE + KE - PE1' - PE2'; Where PE is the potential energy of the molecule before decomposition, KE is the kinetic energy of the molecule before decomposition, PE1' is the potential energy of the molecule after decomposition, and PE2' is the potential energy of the molecule after decomposition; If tmp1+E B ≥0, then exit step S557, otherwise update the kinetic energy KE of the molecule and proceed to the next step S557, and update the kinetic energy KE of the new molecule specifically as follows: Take a random number from 0 to 1 and assign it to q; If tmp1 ≥ 0, the kinetic energy KE1' of the new molecule 1 is: KE1' = tmp1*q; The kinetic energy KE2' of the new molecule 2 is: KE2' = tmp1*(1-q); If tmp1<0, the kinetic energy KE1' and KE2' of the new molecule are obtained by the following formula: KE1'=(tmp1+E B )*q; KE2'=(tmp1+E B -KE1')*q。 8. The method for planning a human-machine collaborative production task according to claim 3, characterized in that: The kinetic energy KE of the new molecule in the single molecule invalid collision reaction in step S56 is specifically calculated as: The temporary variable tmp2 is obtained by the following formula: tmp2 = PE + KE - PE'; Where PE is the potential energy of the molecule before the collision, KE is the kinetic energy of the molecule before the collision, and PE' is the potential energy of the molecule after the collision; If tmp2<0, then exit step S56; If tmp2 ≥ 0, perform the following operations: From L R From 1 to 1, take a random number and assign it to q, then: The kinetic energy KE of the new molecule is tmp2*q; Buffer energy E B Increase to tmp2*(1-q).

9. The method for planning a human-machine collaborative production task according to claim 4, characterized in that: The kinetic energy KE of the new molecule in the multi-molecule molecular synthesis reaction in step S58 is specifically calculated as: The temporary variable tmp3 is obtained by the following formula: tmp3 = PE1 + KE1 + PE2 + KE2 - PE'; Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, and PE' is the potential energy of the new molecule obtained after synthesis; If tmp3>0, the kinetic energy of the new molecule KE=tmp3, otherwise tmp3≤0, the kinetic energy of the new molecule KE=0.

10. The method for planning a human-machine collaborative production task according to claim 5, characterized in that: The kinetic energy KE of the new molecule in the intermolecular collision reaction of multiple molecules in step S59 is specifically calculated as: The temporary variable tmp4 is obtained by the following formula: tmp4=PE1+KE1+PE2+KE2-PE1'-PE2'; Where PE1 is the potential energy of parent molecule 1, PE2 is the potential energy of parent molecule 2, KE1 is the kinetic energy of parent molecule 1, KE2 is the kinetic energy of parent molecule 2, PE1' is the potential energy of the new molecule 1; PE2' is the potential energy of the new molecule 2; If tmp4<0, then exit the current step S59; If tmp4 ≥ 0, perform the following operations: Random numbers from 0 to 1 are assigned to q. The kinetic energy of the new molecule 1 is tmp4*q The kinetic energy of the new molecule 2 is assigned tmp4*(1-q).

Images