CN116307148A - Super heuristic optimization method and system for energy-saving scheduling of hot rolling of steel and product transportation - Google Patents

Abstract

The invention discloses a super heuristic optimization method and a system for energy-saving scheduling of steel hot rolling and product transportation. The invention provides a model of an energy-saving scheduling problem of steel hot rolling and product transportation, and designs a super heuristic algorithm based on artificial bee colony to solve the problem: according to the production stage problem characteristics, a reverse scheduling method is designed to improve the solving efficiency of an algorithm; the quality of the dispatching solution is optimized by the artificial bee colony algorithm combined with the self-adaptive strategy, and the self-adaptive searching strategy guides the algorithm to select the most effective neighborhood operator to operate according to the history searching experience, so that the depth searching capability of the algorithm is improved. The method can obtain a high-quality scheduling scheme for the energy-saving scheduling problem of steel hot rolling and product transportation in a short time, reduce the production and transportation cost of enterprises and improve the hot rolling production benefit of steel enterprises.

Description

Super heuristic optimization method and system for energy-saving scheduling of hot rolling of steel and product transportation

Technical Field

The invention relates to a super heuristic optimization method and a system for energy-saving dispatching of steel hot rolling and product transportation, belonging to the fields of steel production technology and energy-saving dispatching.

Background

The iron and steel industry is indispensable for any industry economy and provides the most important raw materials for other industries. The steel production is used as a prop industry of national economy, and the optimization of the production process has important practical significance for enterprises. The hot rolling stage in the steel production process can be described as a process in which a slab is rolled several times after heating and then edge cut is corrected to a steel sheet, which can be further processed or directly used for production. The hot rolling is used as a key link in the steel production process, and the energy-saving scheduling of hot rolling production and product transportation is reasonably planned, so that the production and transportation energy consumption cost is reduced, and the core competitiveness of steel enterprises is improved.

In the actual hot rolling process of steel, the slab needs to be preheated by a heating furnace before subsequent processing. In the actual production process, hot rolling heating furnace equipment is less, so that the temperature loss of the scheduling method is reduced. The hot rolling process can be abstracted as a mixed flow shop scheduling problem with limited buffer constraints. Under the constraint of a limited buffer zone, the workpiece finishes processing in the previous stage and has no idle processing equipment in the next stage, so that the workpiece preferentially occupies the idle buffer zone to wait for idle machines, otherwise occupies the current processing equipment to wait, and therefore, in order to reduce the production energy consumption cost, the maximum finishing time should be minimized as an optimization target.

CN202110678771. X discloses a method for determining a scheduling plan of hot rolling production of steel, which substitutes the time-of-use electricity price and the adjustable capacity of hot rolling production load in an optimized period into a pre-established scheduling model; CN201611184758.0 discloses an integrated dispatching method for steelmaking-continuous casting-hot rolling, which mainly solves the integrated dispatching problems of steelmaking-continuous casting-hot rolling and stock optimization; but lacks a solution for an integrated scheduling method for hot rolling production and energy saving transportation.

Disclosure of Invention

The invention provides a super heuristic optimization method and a system for energy-saving scheduling of steel hot rolling and product transportation, which are used for obtaining a high-quality solution of steel hot rolling scheduling and energy-saving transportation.

The technical scheme of the invention is as follows:

according to one aspect of the invention, there is provided a super heuristic optimization method for energy-saving scheduling of hot rolling of steel and product transportation, comprising:

taking the minimized cost as an optimization target, and establishing a steel hot rolling production sequencing model and a product transportation energy-saving scheduling model;

designing a reverse scheduling scheme for the production sequencing model, and calculating the finishing time of each workpiece;

and designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

Assuming that the machine start time and the workpiece processing time are known, the steel hot rolling production sequencing model comprises the following steps:

constraint 1: at least one available device exists for each processing stage;

constraint 2: the workpiece can be processed on only one machine at each processing stage;

constraint 3: each machine can only process one workpiece at most at the same time;

constraint 4: all the workpieces have the same processing procedure and are processed strictly according to the sequence of the processing procedures;

constraint 5: after the workpiece is processed in the current processing stage, entering a buffer device to wait until the workpiece is processed in the next stage or directly processed in the next stage, otherwise, occupying the current machine device, waiting for the buffer device until the workpiece is processed in the next stage or waiting for the machine in the next processing stage to be idle for processing;

constraint 6: at least 0 buffer devices are present between two successive processing stages;

constraint 7: the workpiece processing is not allowed to be interrupted and preempted.

Assuming constant vehicle travel speed, the product transportation energy-saving scheduling model is as follows:

h _uvw ≤Q(u,v∈{0,2,...,N _w },u≠v,w∈{1,2,...,m})， (3)

wherein Z is the total cost; θ is a production processing cost coefficient; t represents the finishing time of all workpieces; m represents the total number of vehicles; f is the fixed cost of the vehicle; beta represents a vehicle departure time penalty coefficient; t is t _w The departure time of the vehicle w is the finishing time of the workpiece finally loaded in the vehicle w; alpha represents a vehicle load cost coefficient; n (N) _w The number of clients served by the vehicle w is represented, and u and v are taken as 0 to represent a factory; x is x _uvw As a decision variable, the vehicle w starts from the customer u and reaches the customer v, and is 1, otherwise, is 0; d, d _uv Representing the distance from client u to client v; n represents the total number of clients; h is a _uvw Representing the load capacity of the vehicle w from customer/plant u to customer/plant v; q is the rated load capacity of the vehicle; the objective function and the constraints mean: formula (1) represents minimizing the total cost; formula (2) indicates that all the workpieces are scheduled for delivery; formula (3) indicates that the vehicle load does not exceed its rated load capacity; equations (4) and (5) indicate that the vehicle must return to the factory after all customer service is completed from the factory.

The finishing time of each workpiece is taken as a step 1, and the step 1 specifically comprises the following steps: coding the workpieces in a permutation mode, and calculating the finishing time of each workpiece by adopting a reverse scheduling scheme; in the first stage of the machining process, the work is machined in accordance with a given work machining sequence, and in the subsequent machining stages, the work machining sequence is dynamically arranged in accordance with the work finishing time of the previous stage, the occupation of the intermediate buffer zone and the available machine conditions of the next stage.

The reverse scheduling scheme includes:

the machining time of the workpiece j on the machine k is denoted as p _j,k ，s _i,j,k Representing the start time of the work j on machine k at stage i, c _i,j,k Representing the time at which machine k releases workpiece j in stage i; b _i,j,l The time for workpiece j to enter the first buffer device between the i-th stage and the i+1-th stage; e, e _i,j,l For the time when the workpiece j is released in the buffer device l between the phase I and the phase i+1, and initializing the workpiece numbers in all machines and the buffer device to be-1, and all times to be 0, and initializing i=i-1, the completion time of each workpiece is calculated as follows:

step 1.1: calculating finishing time of all workpieces in the backward direction; recording the earliest released workpiece j in stage i+1 ₁ Where the working machine is k ₁ The earliest occupied buffer device l ₁ The workpiece in (a) is j ₂ Work j finished at earliest in stage i ₃ The machine of (1) is k ₂ The method comprises the steps of carrying out a first treatment on the surface of the If i>0 executing the step 1.2, otherwise executing the step 1.3;

step 1.2: if a buffer device exists between the i and i+1 stages, updating:

if no buffer device exists between the i and i+1 stages, updating:

let i=i-1 and return to step 1.1;

step 1.3: if i=0, selecting a workpiece to process by the earliest idle machine in the 1 st stage according to a given processing sequence;

step 1.4: repeating the steps until all the workpieces are processed.

The design of the ultra heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model solves the problems of hot rolling of steel and energy-saving scheduling of product transportation, and comprises the following steps:

step 2: optimizing the energy-saving scheduling problems of steel hot rolling and product transportation by adopting a hyper-heuristic algorithm; recording a high-level strategy domain as a plurality of low-level heuristic neighborhood operator arrangements, wherein low-level individual codes are randomly generated 1-stage workpiece processing sequence pi, distributing workpieces to each vehicle according to vehicle load constraint and workpiece finishing sequence, generating an initial distribution sequence of each vehicle by adopting a neighbor strategy, and recording key workpieces, namely the workpieces which finish processing in the vehicle at last; the lower-layer individual sequentially executes heuristic operators in the higher-layer individual by adopting a greedy strategy to search;

step 3: and the high-level strategy domain is optimized by adopting an artificial bee colony algorithm to obtain a global optimal solution.

The high-level strategy domain is optimized by adopting an artificial bee colony algorithm to obtain a global optimal solution, and the method comprises the following steps:

step 3.1: initializing a population, namely generating PS (PS) different individuals in a random initialization mode to form an initial population with discrete distribution;

step 3.2: employment of a bee search; the number of hiring bees is the same as the number of population individuals, and each hiring bee randomly selects one individual to perform local search operation and updates the globally optimal solution;

step 3.3: observing bee search; selecting an optimal individual from m individuals by adopting a roulette manner, performing local search operation on the selected individual, and updating a global optimal solution;

step 3.4: searching the scout bees; randomly generating a new individual to replace an original individual which is not updated for the longest time, and executing local search operation on the new individual, and simultaneously updating a global optimal solution;

step 3.5: repeating the steps 3.2-3.4, setting the termination condition of the artificial bee colony algorithm as the maximum iteration number, and outputting the obtained global optimal solution if the iteration number of the algorithm reaches the termination condition.

In the optimization process of the high-level strategy domain by adopting the artificial bee colony algorithm, the termination condition of the local search operation is the maximum number of local search operations, when the local search operation is executed, the new solution improves the original solution, the invalid number of times of the current local search operator is not increased, and otherwise, the invalid number of times of the current local search operator is increased by 1; the next time the local search operator with the smallest invalidation times is preferentially adopted.

The low-level heuristic neighborhood operator is specifically as follows:

l1: in the same vehicle, two workpieces are randomly selected to be sequentially exchanged;

l2: selecting a workpiece to insert into a random position in the same vehicle;

l3: in the same vehicle, selecting to reverse a section of workpiece distribution sequence;

l4: selecting adjacent workpieces and exchanging sequences in the same vehicle;

l5: selecting one workpiece for exchanging among different vehicles, and updating key workpieces;

l6: selecting one workpiece from among different vehicles, randomly inserting the workpiece into another vehicle, and updating the key workpiece;

l7: a machining sequence exchanging operation, namely randomly selecting two workpieces and exchanging machining sequences of the two workpieces;

l8: a machining sequence forward insertion operation for selecting a workpiece forward insertion;

l9: the reverse sequence operation of the processing sequence, selecting a certain section of workpiece sequence and reversing the processing sequence;

l10: the adjacent workpieces of the processing sequence are exchanged, and the adjacent workpieces are selected to exchange the processing sequence.

According to another aspect of the present invention, there is also provided a super heuristic optimization system for hot rolling of steel and energy saving scheduling of product transportation, comprising:

the building module is used for building a steel hot rolling production ordering model and a product transportation energy-saving scheduling model by taking the minimized cost as an optimization target;

the calculation module is used for designing a reverse scheduling scheme for the production sequencing model and calculating the finishing time of each workpiece;

and the solving module is used for designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

The beneficial effects of the invention are as follows: the invention provides a model of an energy-saving scheduling problem of steel hot rolling and product transportation, and designs a super heuristic algorithm based on artificial bee colony to solve the problem: according to the production stage problem characteristics, a reverse scheduling method is designed to improve the solving efficiency of an algorithm; the quality of the dispatching solution is optimized by the artificial bee colony algorithm combined with the self-adaptive strategy, and the self-adaptive searching strategy guides the algorithm to select the most effective neighborhood operator to operate according to the history searching experience, so that the depth searching capability of the algorithm is improved. The method can obtain a high-quality scheduling scheme for the energy-saving scheduling problem of steel hot rolling and product transportation in a short time, reduce the production and transportation cost of enterprises and improve the hot rolling production benefit of steel enterprises.

Drawings

FIG. 1 is a flowchart of an algorithm;

FIG. 2 is a schematic diagram of the hot rolling scheduling problem of steel in the limited buffer zone of the present invention;

FIG. 3 is a schematic diagram of the transportation of the product of the present invention;

FIG. 4 is a schematic diagram of heuristic L1;

FIG. 5 is a schematic diagram of heuristic L2;

FIG. 6 is a schematic diagram of heuristic L3;

FIG. 7 is a schematic diagram of heuristic L4;

FIG. 8 is a schematic diagram of heuristic L5;

FIG. 9 is a schematic diagram of heuristic L6;

FIG. 10 is a schematic diagram of heuristic L7;

FIG. 11 is a schematic diagram of heuristic L8;

FIG. 12 is a schematic diagram of heuristic L9;

fig. 13 is a schematic diagram of heuristic L10.

Detailed Description

The invention will be further described with reference to the drawings and examples, but the invention is not limited to the scope.

Example 1: as shown in fig. 1-13, a super heuristic optimization method for energy-saving scheduling of hot rolling of steel and product transportation comprises the following steps:

taking the minimized cost as an optimization target, and establishing a steel hot rolling production sequencing model and a product transportation energy-saving scheduling model;

designing a reverse scheduling scheme for the production sequencing model, calculating the finishing time of each workpiece and planning a vehicle loading and transporting scheme;

and designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

Further, assuming that the machine start-up time and the workpiece processing time are known, the steel hot rolling production sequencing model is as follows:

constraint 1: at least one available device exists for each processing stage;

constraint 2: the workpiece can be processed on only one machine at each processing stage;

constraint 3: each machine can only process one workpiece at most at the same time;

constraint 4: all the workpieces have the same processing procedure and are processed strictly according to the sequence of the processing procedures; for example, fig. 2 has two stages, there are 4 workpieces, all of which have the same processing procedure, i.e. all of which are processed according to the processing procedures of stage 1 and stage 2, i.e. workpiece 1 completes processing in stage 1 before entering stage 2 to complete processing, and the other workpieces are similar;

constraint 5: after the workpiece is processed in the current processing stage, entering a buffer device to wait until the workpiece is processed in the next stage or directly processed in the next stage, otherwise, occupying the current machine device, waiting for the buffer device until the workpiece is processed in the next stage or waiting for the machine in the next processing stage to be idle for processing;

constraint 6: at least 0 buffers exist between two successive processing stages; the buffer area can be provided with 1 or more buffer devices;

constraint 7: the workpiece processing process does not allow interruption and preemption;

for example, in fig. 2, the processing sequence is stage 1 and stage 2; a work sequence of {1,2,3,4} initial machining sequence, in stage 1, work 1 is machined on machine M1, work 2 is machined on machine M2, then work 3 is allocated, work 4 is allocated; finishing the workpiece 2 in the stage 1, and then directly entering the stage 2 for processing; the 2 nd of the work piece 1 in the stage 1 is processed, and after the work piece 1 is finished, the buffer equipment is empty due to occupation of the stage 2, so that the work piece directly enters the buffer equipment to wait until the work piece enters the stage 2 to be processed; stage 1 workpiece 3, 3 rd finish, and after workpiece 3 finish, since buffer equipment, stage 2, is occupied, current equipment M is occupied ₂ After occupying the current equipment, selecting buffer equipment and entering the first idle stage in the stage 2 (directly entering the stage 2 if the buffer equipment is idle at the same time) until the workpiece 1 finishes processing in the stage 2 (the occupied area is a shaded area in the figure), wherein the workpiece 3 waits for the buffer equipment until the workpiece 1 enters the next stage to process; after the work piece 4 of stage 1 is finished, since the buffer device and stage 2 are occupied, the work piece 4 occupies the current device M ₁ (occupied area is hatched area in the figure), the work 4 waits for the buffer device until, in accordance with the judgment principle of the work 3Entering the next stage of processing; by applying the technical scheme, when idle equipment is selected to enter in the current stage, the priority of the machine in the next stage is selected to be higher than that of the buffer equipment; the workpiece in the buffer device can be taken out at any time;

further, assuming that the vehicle running speed is constant, the product transportation energy-saving scheduling model is as follows:

h _uvw ≤Q(u,v∈{0,2,...,N _w },u≠v,w∈{1,2,...,m})

wherein Z is the total cost; θ is a production processing cost coefficient; t represents the finishing time of all workpieces; m represents the total number of vehicles; f is the fixed cost of the vehicle; beta represents a vehicle departure time penalty coefficient; t is t _w The departure time of the vehicle w is the finishing time of the workpiece finally loaded in the vehicle w; alpha represents a vehicle load cost coefficient; n (N) _w The number of clients served by the vehicle w is represented, and u and v are taken as 0 to represent a factory; x is x _uvw For decision variables, the vehicle w arrives from the customer (or plant) u to the customer (or plant) v with 1, otherwise with 0; d, d _uv Representing the distance from client u to client v; n represents the total number of clients; h is a _uvw Representing the vehicle w from customer (or factory) u to guestLoad capacity at home (or factory) v time; q is the rated load capacity of the vehicle.

Further, the steps of calculating the finishing time of each workpiece and solving the energy-saving scheduling problem of steel hot rolling and product transportation by adopting a super heuristic algorithm based on artificial bee colony are as follows:

step 1: coding the workpieces in a permutation mode, and calculating the finishing time of each workpiece in a reverse scheduling mode; in the first stage of the processing procedure, processing according to a given workpiece processing sequence, and in the subsequent processing stage, dynamically arranging the workpiece processing sequence according to the workpiece finishing time of the previous stage, the occupation condition of an intermediate buffer zone and the available machine condition of the next stage;

further, the step 1 includes:

the machining time of the workpiece j on the machine k is denoted as p _j,k ，s _i,j,k Representing the start time of the work j on machine k at stage i, c _i,j,k Representing the time at which machine k releases workpiece j in stage i; b _i,j,l The time for workpiece j to enter the first buffer device between the i-th stage and the i+1-th stage; e, e _i,j,l For the time when the workpiece j is released in the buffer device l between the phase I and the phase i+1, and initializing all machines and the workpieces in the buffer device to be-1 (virtual workpiece), and all times to be 0, and initializing i=i-1, the completion time of each workpiece is calculated as follows:

step 1.1: calculating finishing time of all workpieces in the backward direction; recording the earliest released workpiece j in stage i+1 ₁ Where the working machine is k ₁ The earliest occupied buffer device l ₁ The workpiece in (a) is j ₂ Work j finished at earliest in stage i ₃ The machine of (1) is k ₂ The method comprises the steps of carrying out a first treatment on the surface of the If i>0 executing the step 1.2, otherwise executing the step 1.3;

step 1.2: if a buffer device exists between the i and i+1 stages, updating:

if no buffer device exists between the i and i+1 stages, updating:

let i=i-1 and return to step 1.1;

step 1.3: if i=0, selecting a workpiece to process by the earliest idle machine in the 1 st stage according to a given processing sequence; the 1 st stage refers to the 1 st process of the processing steps;

step 1.4: repeating the steps until all the workpieces are processed. FIG. 2 shows an example of decoding of a workpiece sequence with an initial machining sequence {1,2,3,4}, first the workpieces 1,2 respectively select the machine M in stage 1 ₁ And M ₂ The workpiece thereafter selects, at each stage, the machine that enters the idle state earliest at each stage for processing (taking workpiece 3 as an example, the machine M that enters the idle state earliest at stage 1 is selected in stage 1) ₂ Processing) if no machine is idle in the next stage, entering a buffer area to wait or occupy the current machine, and finallyResulting in the finishing order 2,1,3,4.

Step 2: optimizing the energy-saving scheduling problems of steel hot rolling and product transportation by adopting a hyper-heuristic algorithm; the high-level strategy domain is formed by arranging 10 low-level heuristic neighborhood operators, the low-level individual codes are randomly generated 1-stage workpiece processing sequence pi, the workpieces are distributed to each vehicle according to the vehicle load constraint and the workpiece finishing sequence, the initial distribution sequence of each vehicle is generated by adopting a neighbor strategy, and meanwhile, key workpieces, namely the workpieces which finish processing in the vehicle at last, are recorded; the lower-layer individual sequentially executes heuristic operators in the higher-layer individual by adopting a greedy strategy to search;

step 3: and the high-level strategy domain is optimized by adopting an artificial bee colony algorithm to obtain a global optimal solution.

Further, the step 3 includes:

step 3.1: initializing a population, namely generating PS (PS) different individuals in a random initialization mode to form an initial population with discrete distribution;

step 3.2: employment of a bee search; the number of hiring bees is the same as the number of population individuals, and each hiring bee randomly selects one individual to perform local search operation and updates the globally optimal solution;

step 3.3: observing bee search; selecting an optimal individual from m individuals by adopting a roulette manner, performing local search operation on the selected individual, and updating a global optimal solution;

step 3.4: searching the scout bees; randomly generating a new individual to replace an original individual which is not updated for the longest time, and executing local search operation on the new individual, and simultaneously updating a global optimal solution; the algorithm is prevented from being sunk into local optimum prematurely;

step 3.5: repeating the steps 3.2-3.4, setting the termination condition of the artificial bee colony algorithm as the maximum iteration number, and outputting the obtained global optimal solution if the iteration number of the algorithm reaches the termination condition.

Further, in the optimization process of the high-level policy domain by adopting the artificial bee colony algorithm, the termination condition of the local search operation is the maximum number of local search operations, when the local search operation is executed, the new solution improves the original solution, the invalid number of times of the current local search operator is not increased, otherwise, the invalid number of times of the current local search operator is increased by 1; the local search operator with the minimum invalid times is preferentially adopted next time; for example, adopting neighborhood operators including insert (insert), multi-swap (multi-swap) and reverse (reverse), and recording the invalid operation times of each local search operator as T_insert, T_multi-swap and T_reverse respectively; when the local search operation is performed, the new solution improves the original solution, then t_insert=0 is recorded, otherwise t_insert=t_insert+1 is recorded; the algorithm preferentially adopts the operation with the minimum invalid times;

further, as shown in fig. 4-13, the low-level heuristic neighborhood operator is specifically as follows:

l1: in the same vehicle, two workpieces are randomly selected to be sequentially exchanged;

l2: selecting a workpiece to insert into a random position in the same vehicle;

l3: in the same vehicle, selecting to reverse a section of workpiece distribution sequence;

l4: selecting adjacent workpieces and exchanging sequences in the same vehicle;

l5: selecting one workpiece for exchanging among different vehicles, and updating key workpieces;

l6: selecting one workpiece from among different vehicles, randomly inserting the workpiece into another vehicle, and updating the key workpiece;

l7: a machining sequence exchanging operation, namely randomly selecting two workpieces and exchanging machining sequences of the two workpieces;

l8: a machining sequence forward insertion operation for selecting a workpiece forward insertion;

l9: the reverse sequence operation of the processing sequence, selecting a certain section of workpiece sequence and reversing the processing sequence;

l10: the adjacent workpieces of the processing sequence are exchanged, and the adjacent workpieces are selected to exchange the processing sequence.

The method can be specifically set up as follows: the population size was set to ps=10, and the number of times of no update limit=100.

To verify the effectiveness of the method of the present invention, table 1 shows the results of the operation of the method of the present invention (hh_abc) and the variable neighborhood search algorithm (VNS) on solving scheduling problems of different problem sizes (S represents the number of processing stages and J represents the number of workpieces). Each algorithm repeatedly solves each example for 20 times in the same time, and MIN, MAX, AVG represents the minimum value, the maximum value and the average value in the 20 solving results respectively.

As can be seen from Table 1, the method of the present invention is superior to the variable neighborhood search algorithm in terms of the solving results of each example, and the effectiveness of the method of the present invention is verified. The scheduling scheme can obtain satisfactory solution of the steel hot rolling scheduling problem constrained by the limited buffer zone in a short time, thereby improving the benefit of hot rolling production of steel enterprises and effectively reducing the production and transportation energy consumption cost.

Example 2: a super heuristic optimization system for energy-efficient scheduling of hot rolling of steel and product transportation, comprising:

the building module is used for building a steel hot rolling production ordering model and a product transportation energy-saving scheduling model by taking the minimized cost as an optimization target;

the calculation module is used for designing a reverse scheduling scheme for the production sequencing model and calculating the finishing time of each workpiece;

and the solving module is used for designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.

In the foregoing embodiments of the present invention, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.

While the present invention has been described in detail with reference to the drawings, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the knowledge of those skilled in the art.

Claims (10)

1. The super heuristic optimization method for energy-saving scheduling of hot rolling of steel and product transportation is characterized by comprising the following steps:

taking the minimized cost as an optimization target, and establishing a steel hot rolling production sequencing model and a product transportation energy-saving scheduling model;

designing a reverse scheduling scheme for the production sequencing model, and calculating the finishing time of each workpiece;

and designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

2. The method for super heuristic optimization of energy-efficient scheduling of hot-rolled steel and product transportation according to claim 1, wherein the hot-rolled steel production ranking model comprises, assuming that the machine start-up time, the work piece processing time are known:

constraint 1: at least one available device exists for each processing stage;

constraint 2: the workpiece can be processed on only one machine at each processing stage;

constraint 3: each machine can only process one workpiece at most at the same time;

constraint 4: all the workpieces have the same processing procedure and are processed strictly according to the sequence of the processing procedures;

constraint 5: after the workpiece is processed in the current processing stage, entering a buffer device to wait until the workpiece is processed in the next stage or directly processed in the next stage, otherwise, occupying the current machine device, waiting for the buffer device until the workpiece is processed in the next stage or waiting for the machine in the next processing stage to be idle for processing;

constraint 6: at least 0 buffer devices are present between two successive processing stages;

constraint 7: the workpiece processing is not allowed to be interrupted and preempted.

3. The method for super heuristic optimization of hot rolling of steel and energy-saving scheduling of product transportation according to claim 1, characterized in that, assuming constant vehicle running speed, the energy-saving scheduling model of product transportation is as follows:

h _uvw ≤Q(u,v∈{0,2,...,N _w },u≠v,w∈{1,2,...,m})，(3)

wherein Z is the total cost; θ is a production processing cost coefficient; t represents the finishing time of all workpieces; m represents the total number of vehicles; f is the fixed cost of the vehicle; beta represents a vehicle departure time penalty coefficient; t is t _w The departure time of the vehicle w is the finishing time of the workpiece finally loaded in the vehicle w; alpha represents a vehicle load cost coefficient; n (N) _w The number of clients served by the vehicle w is represented, and u and v are taken as 0 to represent a factory; x is x _uvw As a decision variable, the vehicle w starts from the customer u and reaches the customer v, and is 1, otherwise, is 0; d, d _uv Representing the distance from client u to client v; n represents the total number of clients; h is a _uvw Representing the load capacity of the vehicle w from customer/plant u to customer/plant v; q is the rated load capacity of the vehicle; the objective function and the constraints mean: formula (1) represents minimizing the total cost; watch (2)Showing that all workpieces are scheduled for delivery; formula (3) indicates that the vehicle load does not exceed its rated load capacity; equations (4) and (5) indicate that the vehicle must return to the factory after all customer service is completed from the factory.

4. The super heuristic optimization method for energy-saving scheduling of hot rolling of steel and transportation of products according to claim 1, wherein the workpieces are coded in a permutation way and a reverse scheduling scheme is adopted to calculate the finishing time of each workpiece; in the first stage of the machining process, the work is machined in accordance with a given work machining sequence, and in the subsequent machining stages, the work machining sequence is dynamically arranged in accordance with the work finishing time of the previous stage, the occupation of the intermediate buffer zone and the available machine conditions of the next stage.

5. The method for super heuristic optimization of energy-efficient scheduling of hot rolling of steel and transportation of products according to claim 1, characterized in that said reverse scheduling scheme comprises:

the machining time of the workpiece j on the machine k is denoted as p _j,k ，s _i,j,k Representing the start time of the work j on machine k at stage i, c _i,j,k Representing the time at which machine k releases workpiece j in stage i; b _i,j,l The time for workpiece j to enter the first buffer device between the i-th stage and the i+1-th stage; e, e _i,j,l For the time when the workpiece j is released in the buffer device l between the phase I and the phase i+1, and initializing the workpiece numbers in all machines and the buffer device to be-1, and all times to be 0, and initializing i=i-1, the completion time of each workpiece is calculated as follows:

step 1.1: calculating finishing time of all workpieces in the backward direction; recording the earliest released workpiece j in stage i+1 ₁ Where the working machine is k ₁ The earliest occupied buffer device l ₁ The workpiece in (a) is j ₂ Work j finished at earliest in stage i ₃ The machine of (1) is k ₂ The method comprises the steps of carrying out a first treatment on the surface of the If i>0 executing the step 1.2, otherwise executing the step 1.3;

step 1.2: if a buffer device exists between the i and i+1 stages, updating:

if no buffer device exists between the i and i+1 stages, updating:

let i=i-1 and return to step 1.1;

step 1.3: if i=0, selecting a workpiece to process by the earliest idle machine in the 1 st stage according to a given processing sequence;

step 1.4: repeating the steps until all the workpieces are processed.

6. The method for super-heuristic optimization of energy-saving scheduling for hot rolling of steel and transportation of products according to claim 1, wherein designing a super-heuristic optimization algorithm based on artificial bee colony according to the energy-saving scheduling model for transportation of products solves the energy-saving scheduling problem for hot rolling of steel and transportation of products, comprising:

step 2: optimizing the energy-saving scheduling problems of steel hot rolling and product transportation by adopting a hyper-heuristic algorithm; recording a high-level strategy domain as a plurality of low-level heuristic neighborhood operator arrangements, wherein low-level individual codes are randomly generated 1-stage workpiece processing sequence pi, distributing workpieces to each vehicle according to vehicle load constraint and workpiece finishing sequence, generating an initial distribution sequence of each vehicle by adopting a neighbor strategy, and recording key workpieces, namely the workpieces which finish processing in the vehicle at last; the lower-layer individual sequentially executes heuristic operators in the higher-layer individual by adopting a greedy strategy to search;

step 3: and the high-level strategy domain is optimized by adopting an artificial bee colony algorithm to obtain a global optimal solution.

7. The super heuristic optimization method for energy-saving scheduling of hot rolling of steel and product transportation according to claim 6, wherein the high-level policy domain adopts a manual bee colony algorithm for optimization to obtain a global optimal solution, comprising:

step 3.1: initializing a population, namely generating PS (PS) different individuals in a random initialization mode to form an initial population with discrete distribution;

step 3.2: employment of a bee search; the number of hiring bees is the same as the number of population individuals, and each hiring bee randomly selects one individual to perform local search operation and updates the globally optimal solution;

step 3.3: observing bee search; selecting an optimal individual from m individuals by adopting a roulette manner, performing local search operation on the selected individual, and updating a global optimal solution;

step 3.4: searching the scout bees; randomly generating a new individual to replace an original individual which is not updated for the longest time, and executing local search operation on the new individual, and simultaneously updating a global optimal solution;

step 3.5: repeating the steps 3.2-3.4, setting the termination condition of the artificial bee colony algorithm as the maximum iteration number, and outputting the obtained global optimal solution if the iteration number of the algorithm reaches the termination condition.

8. The super heuristic optimization method for steel hot rolling and product transportation energy saving scheduling according to claim 7, wherein in the optimization process of the high-level strategy domain by adopting a manual bee colony algorithm, the termination condition of the local search operation is the maximum number of local search operations, when the local search operation is executed, the new solution improves the original solution, the invalid number of the current local search operator is not increased, otherwise, the invalid number of the current local search operator is increased by 1; the next time the local search operator with the smallest invalidation times is preferentially adopted.

9. The super heuristic optimization method for steel hot rolling and product transportation energy saving scheduling according to claim 6, wherein the low-level heuristic neighborhood operators are specifically as follows:

l1: in the same vehicle, two workpieces are randomly selected to be sequentially exchanged;

l2: selecting a workpiece to insert into a random position in the same vehicle;

l3: in the same vehicle, selecting to reverse a section of workpiece distribution sequence;

l4: selecting adjacent workpieces and exchanging sequences in the same vehicle;

l5: selecting one workpiece for exchanging among different vehicles, and updating key workpieces;

l6: selecting one workpiece from among different vehicles, randomly inserting the workpiece into another vehicle, and updating the key workpiece;

l7: a machining sequence exchanging operation, namely randomly selecting two workpieces and exchanging machining sequences of the two workpieces;

l8: a machining sequence forward insertion operation for selecting a workpiece forward insertion;

l9: the reverse sequence operation of the processing sequence, selecting a certain section of workpiece sequence and reversing the processing sequence;

l10: the adjacent workpieces of the processing sequence are exchanged, and the adjacent workpieces are selected to exchange the processing sequence.

10. An ultra-heuristic optimization system for energy-saving scheduling of hot rolling of steel and product transportation, which is characterized by comprising:

the building module is used for building a steel hot rolling production ordering model and a product transportation energy-saving scheduling model by taking the minimized cost as an optimization target;

the calculation module is used for designing a reverse scheduling scheme for the production sequencing model and calculating the finishing time of each workpiece;

and the solving module is used for designing a super heuristic optimization algorithm based on artificial bee colony according to the product transportation energy-saving scheduling model to solve the problems of hot rolling of steel and energy-saving scheduling of product transportation.

Images