CN109255512B - University course arrangement method based on Monte Carlo genetic algorithm - Google Patents

Abstract

A university course arrangement method based on Monte Carlo genetic algorithm, carry on splice form a time strip at random to the class, course, teacher, time and classroom information obtained at first, then, form an individual, namely the school timetable, a plurality of individuals form a group by a plurality of time strips that meet constraint condition; selecting an individual with the optimal fitness of the current population, carrying out cross variation on the individual, and adopting a Monte Carlo probability receiving method in the cross process; and finally, transforming according to the set iteration times, and taking the optimal individual in the population of the last generation as a final prediction result. The invention provides a course arrangement method for colleges and universities, which can reasonably allocate various teaching resources.

Description

University course arrangement method based on Monte Carlo genetic algorithm

Technical Field

The invention relates to the fields of course arrangement in teaching, intelligent optimization and computer application in colleges and universities, in particular to a course arrangement method in colleges and universities based on a Monte Carlo genetic algorithm.

Background

In recent years, with the implementation of national talent plans, the demand of colleges and universities for students is increased, and meanwhile, the phenomena of shortage of college resource handling, increased pressure of teaching management and the like occur. Especially in the aspect of course arrangement in colleges and universities, before the course is opened in each school period, the educational administration staff arranges a group of proper teaching time and space, namely a school timetable, for the opened course by utilizing the existing teaching resources, so that students can reasonably arrange the time, the learning efficiency is improved, and the teaching work can normally and orderly run. At present, most institutes and universities still adopt manual course arrangement, with the gradual expansion of the study scale of colleges and universities, the types of subjects, the professional number, the class number, the number of classes and the number of students are increased, the situation that classroom resource conflict or teacher resource conflict is difficult to solve by manual course arrangement is more and more, the inevitable work is complicated, the workload is huge, the defects of low efficiency are more and more prominent, and meanwhile, the manual operation is not easy to fully utilize resources to meet the constantly changing requirements.

The randomness of relying on human brain is strong when people arrange classes manually, and the people do not have strict working steps, think which arranges which, so the people can be out of consideration. The computer works in a mode different from that of the human brain, does not have the divergent thinking ability of the human brain, converts all information into data, converts the human thinking into specific rules and algorithms, and processes the data by using the algorithms through the computer language. If the computer is used for simulating the thinking process of the human brain to carry out course arrangement, the computer has high running speed and high processing capacity, and can quickly obtain a feasible scheme meeting the constraint, so that a plurality of scientific, accurate and used course arrangement schemes are compiled for teachers to select and then are manually finely adjusted, the workload of the teachers can be greatly reduced, the teachers can finish other work within more time, the teaching management work efficiency is improved, various teaching resources are optimized and configured, the teaching management quality of the whole school is also improved, and the informatization process of teaching reform of colleges and universities is promoted.

Disclosure of Invention

In order to overcome the defects that the situation that classroom resource conflict or teacher resource conflict is difficult to solve in manual course arrangement and inevitable work is complicated, the workload is huge and the efficiency is low in course arrangement, the invention provides a university course arrangement method based on a Monte Carlo genetic algorithm.

The technical scheme adopted by the invention for solving the technical problems is as follows:

a university course arrangement method based on a Monte Carlo genetic algorithm comprises the following steps:

1) and (3) encoding: acquiring class basic information C ═ { C ═ C₁,c₂,c₃,…,c_j,…,c_nThe basic information W ═ W₁,w₂,w₃,…,w_h,…,w_mThe teacher basic information T ═ T₁,t₂,t₃,…,t_u,…,t_vThe class time information D ═ D }₁,d₂,d₃,…,d_y,…,d_lAnd classroom basic information P ═ P₁,p₂,p₃,…,p_r,…,p_bWhere n denotes the total number of classes, m denotes the total number of classes, v denotes the total number of teachers, l denotes the total number of hours in class, b denotes the total number of classrooms, c denotes the total number of classrooms_j、w_h、t_u、d_y、p_rThe inner codes are each composed of 4 decimal numbers. For a period of time d_yWhen coding, d_y＝d_y1d_y2d_y3d_y4Dividing a day into five time segments, wherein d_y1Indicating that the course is first on the day of the week, d_y2Indicating the time period during which the lesson was first attended, d_y3Indicating that the course is on the second day of the week, d_y4Indicating the time period during which the lesson is to be taken for the second time, e.g. d_y1225 represents the second lesson day of the week, the first lesson day, and the fifth lesson day of the week. For course w_hWhen coding, each course has a corresponding one-day time period class efficiency value w_h＝{w_h1,w_h2,w_h3,w_h4,w_h5In which w_h1A class efficiency value representing a first time period. For teacher t_uWhen coding, each timeEach teacher has an efficiency value t corresponding to the time period of one day_u＝{t_u1,t_u2,t_u3,t_u4,t_u5}. Through the analysis and classification of class, course, teacher, class time and classroom information, the class, the teacher and the classroom information are finally expressed as A_i＝c_j～w_h～t_u～d_y～p_rForm is coded, wherein A_iThe time bar is called the ith time bar in a population, the time bar represents that j classes teach h courses in r classrooms by u teachers in y time periods, and individuals in the population are formed by j x h time bars;

2) the fitness function of an individual is shown in equation (1),

wherein f is_i＝k₁*f_i1+k₂*f_i2+k₃*f_i3Indicating the fitness of the ith time bar in the individual, wherein

Shows the influence of the interval between two sessions in a week,

representing the impact of the course scheduling during a week, wherein z₁,z₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing a second time period, z₁Then is 2, therefore

Is denoted as w_h2The efficiency value of the air conditioner is improved,

represents the impact of the teacher's schedule over the week, where x₁,x₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing the second time period, x₁Then is 2, therefore

Then it is denoted as t_u2Efficiency value, k₁,k₂,k₃Representing their corresponding weights;

3) population initialization, the process is as follows:

the population size NP is characterized in that the current population algebra stage is 1, the variation probability is mutate, the maximum iteration number is N, and the constraint condition of the population is as follows: a teacher can only arrange one course in the same time period, one class can only arrange one course in the same time period, the same classroom can only arrange one course in the same time period, and the capacity of the classroom should be larger than the number of people in class; randomly generating initial population Q meeting constraint conditions_init＝{S₁,S₂,…,S_NPIn which S is_NPRepresents the Nth class schedule, i.e. the individual;

4) population crossing, the process is as follows:

4.1) calculating the current population Q_currentSelecting the optimal individual S according to the fitness of each individual_best；

4.2) in the optimal Individual S_bestRandomly selecting a time bar S_bestA_randomThen randomly selecting a time bar S from other individuals in the population_otherA_otherrandomThe class time d in the two time bars_yAnd classroom p_rCarrying out exchange;

4.3) in the optimal Individual S_bestJudging whether the exchanged individuals meet constraint conditions, if so, continuing to execute the next step, and if not, returning to the step 4.2);

4.4) calculating the individual fitness S after exchange_bestF_behindAnd the optimal individual fitness S before exchange_bestF_frontDifference between them Δ F_behind-front＝S_bestF_behind-S_bestF_fr_ontIf Δ F_behind-frontIf the number is more than 0, the individuals with better cross yield are received as next generation individuals, other individuals are not changed, and finally the generation is carried outProgeny population Q_reproduct(ii) a If Δ F_behind-front< 0 using the Monte Carlo probability

Receiving, where KT is the temperature coefficient, if Z_MontoIf > rand (0,1), receiving the exchange of the time bar; otherwise, refusing the time bar exchange;

5) population variation, the process is as follows:

calculating newly generated offspring population Q_reproductSelecting the optimal individual S according to the fitness of each individual_{reproductbest}Randomly selecting a time bar S from the optimal individuals_{reproductbest}A_randomIf the mute is more than random (0,1), the class time and the classroom information are randomly selected from the class time D and the classroom information P to re-time the time bar S_{reproductbest}A_randomAssembling the class time and classroom information; otherwise, no operation is performed;

6) updating the current population algebra stage to stage +1, and returning to the step 4 if the stage is less than N; otherwise, the maximum iteration number is met, and the population optimization is finished.

The technical conception of the invention is as follows: firstly, randomly splicing acquired class, course, teacher, time and classroom information to form a time bar, then forming an individual, namely a class schedule, by a plurality of time bars meeting constraint conditions, and forming a group by a plurality of individuals; selecting an individual with the optimal fitness of the current population, carrying out cross variation on the individual, and adopting a Monte Carlo probability receiving method in the cross process; and finally, transforming according to the set iteration times, and taking the optimal individual in the population of the last generation as a final prediction result.

The beneficial effects of the invention are as follows: the Monte Carlo probability receiving method is integrated into the intersection of the genetic algorithm, so that the population can better jump out of the local optimal condition, and the population is closer to the optimal solution.

Drawings

FIG. 1 is a basic flow chart of a monte Carlo genetic algorithm course arrangement method.

Fig. 2 is a school timetable diagram obtained by the monte carlo genetic algorithm course arrangement method.

Detailed Description

The invention is further described below with reference to the accompanying drawings.

Referring to fig. 1 and 2, a university course scheduling method based on a monte carlo genetic algorithm includes the following steps:

1) and (3) encoding: acquiring class basic information C ═ { C ═ C₁,c₂,c₃,…,c_j,…,c_nThe basic information W ═ W₁,w₂,w₃,…,w_h,…,w_mThe teacher basic information T ═ T₁,t₂,t₃,…,t_u,…,t_vThe class time information D ═ D }₁,d₂,d₃,…,d_y,…,d_lAnd classroom basic information P ═ P₁,p₂,p₃,…,p_r,…,p_bWhere n denotes the total number of classes, m denotes the total number of classes, v denotes the total number of teachers, l denotes the total number of hours in class, b denotes the total number of classrooms, c denotes the total number of classrooms_j、w_h、t_u、d_y、p_rThe inner codes are each composed of 4 decimal numbers. For a period of time d_yWhen coding, d_y＝d_y1d_y2d_y3d_y4Dividing a day into five time segments, wherein d_y1Indicating that the course is first on the day of the week, d_y2Indicating the time period during which the lesson was first attended, d_y3Indicating that the course is on the second day of the week, d_y4Indicating the time period during which the lesson is to be taken for the second time, e.g. d_y1225 represents the second lesson day of the week, the first lesson day, and the fifth lesson day of the week. For course w_hWhen coding, each course has a corresponding one-day time period class efficiency value w_h:{w_h1,w_h2,w_h3,w_h4,w_h5In which w_h1A class efficiency value representing a first time period. For teacher t_uWhen coding, each teacher has a corresponding one-day time periodEfficiency value t in class_u:{t_u1,t_u2,t_u3,t_u4,t_u5}. Through the analysis and classification of class, course, teacher, class time and classroom information, the class, the teacher and the classroom information are finally expressed as A_i＝c_j～w_h～t_u～d_y～p_rForm is coded, wherein A_iThe time bar is called the ith time bar in a population, the time bar represents that j classes teach h courses in r classrooms by u teachers in y time periods, and individuals in the population are formed by j x h time bars;

2) the fitness function of an individual is shown in equation (1),

wherein f is_i＝k₁*f_i1+k₂*f_i2+k₃*f_i3Indicating the fitness of the ith time bar in the individual, wherein

Shows the influence of the interval between two sessions in a week,

representing the impact of the course scheduling during a week, wherein z₁,z₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing a second time period, z₁Then is 2, therefore

Is denoted as w_h2The efficiency value of the air conditioner is improved,

represents the impact of the teacher's schedule over the week, where x₁,x₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing the second time period, x₁Then is 2, therefore t_uz1Then watchShown as t_u2Efficiency value, k₁,k₂,k₃Representing their corresponding weights;

3) population initialization, the process is as follows:

the population size NP is characterized in that the current population algebra stage is 1, the variation probability is mutate, the maximum iteration number is N, and the constraint condition of the population is as follows: a teacher can only arrange one course in the same time period, one class can only arrange one course in the same time period, and the capacity of the classroom should be larger than the number of people in the class; randomly generating initial population Q meeting constraint conditions_init＝{S₁,S₂,…,S_NPIn which S is_NPRepresents the Nth class schedule, i.e. the individual;

4) population crossing, the process is as follows:

4.1) calculating the current population Q_currentSelecting the optimal individual S according to the fitness of each individual_best；

4.2) in the optimal Individual S_bestRandomly selecting a time bar S_bestA_randomThen randomly selecting a time bar S from other individuals in the population_otherA_otherrandomThe class time d in the time bar_yAnd classroom p_rCarrying out exchange;

4.3) in the optimal Individual S_bestJudging whether the individual meets the constraint condition in the exchanged time bar, if so, continuing to execute the next step, and if not, returning to the step 4.2);

4.4) calculating the exchanged individual fitness S_bestF_behindAnd the optimal individual fitness S before exchange_bestF_frontDifference between them Δ F_behind-front＝S_bestF_behind-S_bestF_frontIf Δ F_behind-frontIf the number is more than 0, the better individuals are obtained through crossing, the better individuals are received as next generation individuals, other individuals are unchanged, and finally the offspring population Q is generated_reproduct(ii) a If Δ F_behind-front< 0 using the Monte Carlo probability

Receiving, where KT is a temperature coefficient, is a constant, if Z_MontoIf > rand (0,1), receiving the exchange of the time bar; otherwise, refusing the time bar exchange;

5) population variation, the process is as follows:

calculating the current population, namely the newly generated offspring population Q_reproductThe fitness of each individual is summed, and the optimal individual S is selected_{reproductbest}Randomly selecting a time bar S from the optimal individuals_{reproductbest}A_randomIf the mute is more than random (0,1), the class time and the classroom information are randomly selected from the class time D and the classroom information P to re-time the time bar S_{reproductbest}A_randomAssembling the class time and classroom information; otherwise, no operation is performed;

6) updating the current population algebra stage to stage +1, and returning to the step 4 if the stage is less than N; otherwise, the maximum iteration number is met, and the population optimization is finished.

The example takes a college of colleges and universities in Hangzhou as an example, and the college course arrangement method based on the Monte Carlo genetic algorithm comprises the following steps:

1) and (3) encoding: the obtained class, course, teacher, class time and classroom information are coded, wherein the code C of the class is {1101,1102,1103,1104,1105,1106,1107,1108,1109,1110,1111,1112}, the code W of the course is {1101,1102,1103,1104,1105,1106,1107,1108,1109,1110}, the code T of the teacher is {1101,1102,1103,1104,1105,1106,1107,1108,1109,1110}, and the code D of the class time is {1121,1122,1123,1124,1125,1131,1132,1133,1134,1135,1141,1142,1143,1144,1145,1151,1152,1153,1154,1155,1221,1222, 3,1224,1225,1231,1232,1233,1234, 1245, 1241,1242,1243,1244,1245,1251,1252,1253,1254,1255,1321, 1323,1324,1325, 1421, 1332,1333,1334, 1345, 1342,1343, 1353, 1553, 1551, 1553, 1551, 1554, 1553, 1433, 1431, 1433, 1554, 1553,1554, 1553, 1433, 1554, 1553,1554, 1553,1554, 1553, 1432, 1553, 1551, 1553,1554, 2141, 1553,1554, 1553, 1551, 1554, 1553, 1551, 1553,1554, 1553,1554, 1553, 1551, 1554, 1553, 3,1554, 1553,1554, 1553, 1551, 1554, 1551, 1553, 1551, 1553, 3, 1553,1554, 1533, 1554, 1551, 1553,1554, 1551, 1554, 3,1554,15551,2152,2153,2154,2155,2231,2232,2233,2234,2235,2241,2242,2243,2244,2245,2251,2252,2253,2254,2255,2331,2332,2333,2334,2335,2341,2342,2343,2344,2345,2351,2352,2353,2354,2355,2431,2432,2433,2434,2435,2441,2442,2443,2444,2445,2451,2452,2453,2454,2455,2531,2532,2533,2534,2535,2541,2542,2543,2544,2545,2551,2552,2553,2554,2555,3141,3142,3143,3144,3145,3151,3152,3153,3154,3155,3241,3242,3243,3244,3245,3251,3252,3253,3254,3255,3341,3342,3343,3344,3345,3351,3352,3353,3354,3355,3441,3442,3443,3444,3445,3451,3452,3453,3454,3455,3541,3542,3543,3544,3545,3551,3552,3553,3554,3555,4151,4152,4153,4154,4155,4251,4252,4253,4254,4255,4351,4352,4353,4354,4355,4451,4452,4453,4454,4455,4551,4552,4553,4554,4555} and a classroom encoding P ═ 0114,0125,0134,0211,0224,0234,0335,0334,0341, where for 1101 in the class encoding, the first digit represents the corresponding college, the second digit represents the corresponding specialty, the last two digits represent the class, for 1101 in the course encoding, the first digit represents the corresponding college, the second digit represents the corresponding specialty, the last two digits represent the course, for 1101 in the teacher encoding, the first digit represents the corresponding college, the second digit represents the corresponding specialty, the last two digits represent the teacher's job number, and for time period D encoding, the day is divided into five time periods, such as D_y1225 represents the second lesson day of the week, the first lesson day, and the fifth lesson day of the week. In course coding, each course has an efficiency value corresponding to a time of day, such as efficiency value 1101 corresponding to course 1101: {0.5,0.25,0,0.25,0.5}, representing 1101 class benefit values for the class over 5 time periods. In teacher coding, each teacher has efficiency values 1101 corresponding to the time period of one day, for example, the efficiency value 1101 corresponds to the teacher number 1101: {1,0,0.25,0.25,0.5}. Finally, integrating the class, the course, the teacher, the class time and the classroom codes to form a time bar;

2) the fitness function of an individual is shown in equation (1),

wherein f is_i＝k₁*f_i1+k₂*f_i2+k₃*f_i3Indicating the fitness of the ith time bar in the individual, wherein

Shows the influence of the interval between two sessions in a week,

representing the impact of the course scheduling during a week, wherein z₁,z₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing a second time period, z₁Then is 2, therefore

Is denoted as w_h2The efficiency value of the air conditioner is improved,

represents the impact of the teacher's schedule over the week, where x₁,x₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing the second time period, x₁Then is 2, therefore

Then it is denoted as t_u2Efficiency value, where k₁＝0.6,k₂＝0.3,k₃＝0.1；

3) Population initialization, the process is as follows:

the population scale is 50, the current population algebra stage is 1, the variation probability is 0.1, the maximum iteration number is 1000, and the constraint condition of the population is as follows: a teacher can only arrange one course in the same time period, one class can only arrange one course in the same time period, and the capacity of the classroom should be larger than the number of people in the class; randomly generating initial population Q meeting constraint conditions_init＝{S₁,S₂,…,S₅₀}；

4) Population crossing, the process is as follows:

4.1) calculating the current population Q_currentSelecting the optimal individual S according to the fitness of each individual_best；

4.2) in the optimal Individual S_bestRandomly selecting a time bar S_bestA_randomThen randomly selecting a time bar S from other individuals in the population_otherA_otherrandomThe class time d in the time bar_yAnd classroom p_rCarrying out exchange;

4.3) in the optimal Individual S_bestJudging whether the individual meets the constraint condition in the exchanged time bar, if so, continuing to execute the next step, and if not, returning to the step 4.2);

4.4) calculating the exchanged individual fitness S_bestF_behindAnd the optimal individual fitness S before exchange_bestF_frontDifference between them Δ F_behind-front＝S_bestF_behind-S_bestF_frontIf Δ F_behind-frontIf the number is more than 0, the better individuals are obtained through crossing, the better individuals are received as next generation individuals, other individuals are unchanged, and finally the offspring population Q is generated_reproduct(ii) a If Δ F_behind-front< 0 using the Monte Carlo probability

Receiving, where KT ═ 2, is a constant, if Z_MontoIf > rand (0,1), receiving the exchange of the time bar; otherwise, refusing the time bar exchange;

5) population variation, the process is as follows:

calculating the current population, namely the newly generated offspring population Q_reproductThe fitness of each individual is summed, and the optimal individual S is selected_{reproductbest}Randomly selecting a time bar S from the optimal individuals_{reproductbest}A_randomIf the mute is more than random (0,1), the class time and the classroom information are randomly selected from the class time D and the classroom information P to re-time the time bar S_{reproductbest}A_randomAssembling the class time and classroom information; whether or notThen, do not operate;

6) updating the current population algebra stage to stage +1, and returning to the step 4 if the stage is less than N; otherwise, the maximum iteration number is met, and the population optimization is finished.

While the foregoing has described the preferred embodiments of the present invention, it will be apparent that the invention is not limited to the embodiments described, but can be practiced with modification without departing from the essential spirit of the invention and without departing from the spirit of the invention.

Claims (1)

1. A university course arrangement method based on a Monte Carlo genetic algorithm is characterized by comprising the following steps:

1) and (3) encoding: acquiring class basic information C ═ { C ═ C₁,c₂,c₃,…,c_j,…,c_nThe basic information W ═ W₁,w₂,w₃,…,w_h,…,w_mThe teacher basic information T ═ T₁,t₂,t₃,…,t_u,…,t_vThe class time information D ═ D }₁,d₂,d₃,…,d_y,…,d_lAnd classroom basic information P ═ P₁,p₂,p₃,…,p_r,…,p_bWhere n denotes the total number of classes, m denotes the total number of classes, v denotes the total number of teachers, l denotes the total number of hours in class, b denotes the total number of classrooms, c denotes the total number of classrooms_j、w_h、t_u、d_y、p_rThe inner codes are all composed of 4 decimal numbers; for a period of time d_yWhen coding, d_y＝d_y1d_y2d_y3d_y4Dividing a day into five time segments, wherein d_y1Indicating that the course is first on the day of the week, d_y2Indicating the time period during which the lesson was first attended, d_y3Indicating that the course is on the second day of the week, d_y4Indicating the time period during which the lesson is to be taken for the second time, e.g. d_y1225 represents the second lesson of the week with the first lesson time of Monday, the second lessonThe fifth section of the lesson time of tuesday; for course w_hWhen coding, each course has a corresponding one-day time period class efficiency value w_h＝{w_h1,w_h2,w_h3,w_h4,w_h5In which w_h1A class efficiency value representing a first time period; for teacher t_uDuring coding, each teacher has a corresponding class efficiency value t in one day time period_u＝{t_u1,t_u2,t_u3,t_u4,t_u5}; through the analysis and classification of class, course, teacher, class time and classroom information, the class, the teacher and the classroom information are finally expressed as A_i＝c_j～w_h～t_u～d_y～p_rForm is coded, wherein A_iThe time bar is called the ith time bar in a population, the time bar represents that j classes teach h courses in r classrooms by u teachers in y time periods, and individuals in the population are formed by j x h time bars;

2) the fitness function of an individual is shown in equation (1),

wherein f is_i＝k₁*f_i1+k₂*f_i2+k₃*f_i3Indicating the fitness of the ith time bar in the individual, wherein

Shows the influence of the interval between two sessions in a week,

representing the impact of the course scheduling during a week, wherein z₁,z₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing a second time period, z₁Then is 2, therefore

Is denoted as w_h2The efficiency value of the air conditioner is improved,

represents the impact of the teacher's schedule over the week, where x₁,x₂Respectively represents d_y2And d_y4Index of corresponding time period, e.g. if d_y2When representing the second time period, x₁Then is 2, therefore

Then it is denoted as t_u2Efficiency value, k₁,k₂,k₃Representing their corresponding weights;

3) population initialization, the process is as follows:

the population size NP is characterized in that the current population algebra stage is 1, the variation probability is mutate, the maximum iteration number is N, and the constraint condition of the population is as follows: a teacher can only arrange one course in the same time period, one class can only arrange one course in the same time period, the same classroom can only arrange one course in the same time period, and the capacity of the classroom should be larger than the number of people in class; randomly generating initial population Q meeting constraint conditions_init＝{S₁,S₂,…,S_NPIn which S is_NPRepresents the Nth class schedule, i.e. the individual;

4) population crossing, the process is as follows:

4.1) calculating the current population Q_currentSelecting the optimal individual S according to the fitness of each individual_best；

4.2) in the optimal Individual S_bestRandomly selecting a time bar S_bestA_randomThen randomly selecting a time bar S from other individuals in the population_otherA_otherrandomThe class time d in the two time bars_yAnd classroom p_rCarrying out exchange;

4.3) in the optimal Individual S_bestJudging whether the exchanged individuals meet constraint conditions, if so, continuing to execute the next step, and if not, returning to the step 4.2);

4.4) calculating the individual fitness S after exchange_bestF_behindAnd the optimal individual fitness S before exchange_bestF_frontDifference between them Δ F_behind-front＝S_bestF_behind-S_bestF_frontIf Δ F_behind-frontIf the number is more than 0, the better individuals are obtained through crossing, the better individuals are received as next generation individuals, other individuals are unchanged, and finally the offspring population Q is generated_reproduct(ii) a If Δ F_behind-front< 0 using the Monte Carlo probability

Receiving, where KT is the temperature coefficient, if Z_MontoIf > rand (0,1), receiving the exchange of the time bar; otherwise, refusing the time bar exchange;

5) population variation, the process is as follows:

calculating newly generated offspring population Q_reproductSelecting the optimal individual S according to the fitness of each individual_{reproductbest}Randomly selecting a time bar S from the optimal individuals_{reproductbest}A_randomIf the mute is more than random (0,1), the class time and the classroom information are randomly selected from the class time D and the classroom information P to re-time the time bar S_{reproductbest}A_randomAssembling the class time and classroom information; otherwise, no operation is performed;

6) updating the current population algebra stage to stage +1, and returning to the step 4 if the stage is less than N; otherwise, the maximum iteration number is met, and the population optimization is finished.

Images