CN-122022300-A - Steel plate stack position allocation optimization method based on LOT combination and split recombination

CN122022300ACN 122022300 ACN122022300 ACN 122022300ACN-122022300-A

Abstract

The invention discloses a steel plate stack allocation optimization method based on LOT combination and split recombination, which comprises the following steps of S1, collecting steel plate parameters including LOT, size specification and ex-warehouse date of a steel plate, defining a LOT allocation unit, determining stack field parameters, S2, designing a steel plate stack allocation rule and an optimization model, S3, setting genetic algorithm parameters, representing the mapping relation between LOT and stack positions through chromosome coding, and solving through selection, crossing and mutation operations to obtain a preliminary stack position allocation scheme, S4, designing a simulated annealing optimized genetic algorithm, and S5, carrying out second-stage LOT split recombination operation through a hybrid algorithm to obtain a final stack position allocation scheme. The method fully considers information of LOT (processing batch), warehouse-out date and the like of the steel plates, reduces the number of occupied stacking bits on the premise of ensuring zero invalid turning plate operation by reasonably planning the stacking position of the steel plates, and improves the space utilization rate.

Inventors

JING XUWEN
YUE SHUTING
LIU JINFENG
CHEN YU
LI SU
XIE YANG
QI PENGCHENG

Assignees

江苏科技大学

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. The steel plate stack position allocation optimization method based on LOT combination and splitting recombination is characterized by comprising the following steps of: S1, acquiring parameters of a steel plate, including LOT, dimension specification and delivery date of the steel plate, defining a LOT distribution unit, and determining parameters of a storage yard; s2, designing a steel plate stack position distribution rule and an optimization model; s3, setting genetic algorithm parameters, representing the mapping relation between LOT and the stack position through chromosome coding, and solving to obtain a preliminary stack position distribution scheme through selection, crossing and mutation operations; S4, designing a genetic algorithm for simulating annealing optimization; S5, carrying out LOT splitting and recombining operation at the second stage through a mixing algorithm to obtain a final stack position distribution scheme.
2. The method for optimizing stack allocation of steel plates based on LOT combination and split reorganization according to claim 1, wherein in step S1, a LOT of steel plates is used as a core unit for stack allocation, and parameters of a stack field including length x width dimensions of each stack in the stack field and a capacity of steel plates that can be stacked are determined.
3. The method for optimizing stack allocation of steel plates based on LOT combination and split reorganization according to claim 1, wherein in step S2, the stack allocation rule is different LOTs stacked in the same stack, and the LOT with early delivery date is required to be placed above the LOT with late delivery date when stacking, and the optimizing model is the minimum stack number, and the mathematical model is expressed as follows: ; Where j= [1,2, ], m ] is the set of available stack bits, y j is 1 if stack bit J is used at least once, otherwise is 0.
4. A method for optimizing stack allocation of steel plates based on LOT combination and split reorganization according to claim 3, wherein the constraint conditions corresponding to the steel plates in stack allocation include: Each sheet must be assigned a stack, expressed as: ; Wherein i= [1, 2..n ] is a set of steel plates, wherein each steel plate I is uniquely subordinate to one LOT, which represents the kth LOT allocation unit, wherein K e K, k= [1, 2..p ] is a composite LOT, wherein each index corresponds to one unique segment-LOT combination, 1 if the steel plate I is allocated to the stack j, x ij , otherwise 0; each layer of each stack is at most stored with one steel plate expressed as: ; The height of the steel plates stacked in the stack position is not more than the height limit of the stack position, and the steel plates are expressed as follows: ; wherein t i is the thickness of the steel plate i in the subordinate LOTk, and H j is the height of the stack j; The delivery time of the upper steel plate in the stack is earlier than that of the lower steel plate, and is expressed as: ; Wherein, o i is the delivery date of the upper layer steel plate i in the stack position, f ij is the layer number of the upper layer steel plate i in the stack position j, o i' is the delivery date of the lower layer steel plate i 'of the steel plate i in the stack position stack, and f i'j is the layer number of the lower layer steel plate i' in the stack position j; Ensure that If and only if the stack j is used, it is expressed as: ; ensure that if any steel plate in the stack j comes from LOTk M k is the number of steel plates in LOT k: ; Wherein, the G i is the LOT number to which the steel plate i belongs, and z kj represents whether LOT k is allocated to the stack j; ensuring that if LOT k is allocated to multiple stacks, i.e Then : ; Where s k represents whether LOTk is split into multiple stacks.
5. The method for optimizing stack allocation of steel plates based on LOT combination and splitting recombination according to claim 4 is characterized in that in step S3, all steel plate information is input according to the LOT grouping, population initialization and fitness calculation are carried out under constraint conditions and allocation rule conditions, if the optimal result is obtained, the optimal result is directly output, otherwise, selection, crossover and mutation operations are carried out, a new generation of population is generated, and the fitness of the new generation of population is continuously evaluated until the optimal result is found.
6. The optimization method for steel plate stack allocation based on the LOT combination and the split recombination according to claim 5, wherein the genetic algorithm parameters comprise a population size n=50, iteration times m=100, crossover probability crossrate =0.8 and variation probability mutarate =0.15; The coding mode is stack center coding, chromosome C is defined as a list of genes with the length of |J| and the length of the list is equal to the total number m of available stacks, each gene corresponds to a specific stack J, the structure of the gene is a nested list with variable length, and the structure of the gene comprises all LOT information distributed to the stack, wherein the chromosome structure is expressed as: ; wherein, m= |J| and the stack gene structure is expressed as follows: ; Wherein k j represents the LOT number stacked in the stack position j, k j is more than or equal to 0, the list order indicates the stacking sequence, and the index 0 is the bottommost layer; each LOT unit in the list represents a structure, and the LOT information includes the LOT number All steel plates belonging to LOTk Stack level of LOT Latest date of delivery of all steel sheets in LOT ; The entire solution can be represented as a hierarchy: 。
7. The optimization method for stack allocation of steel plates based on LOT combination and split recombination according to claim 5, wherein in the cross operation, the excellent local stacking mode in the parent is inherited by exchanging allocation content of the whole stack between the chromosomes of the parent, and if individuals which do not meet the height constraint are generated after the cross operation, a penalty function method is adopted to process the constraint; In the mutation operation, a non-empty LOT unit k is randomly selected from a chromosome, a current stack gene J is recorded, another target stack index J ' is randomly selected from a stack set J, the selected LOT unit k is removed from a LOT list contained in the stack J and is inserted into a random position in the stack gene J ' list, after the mutation operation, the target stack J ' is checked according to the delivery date and the stack height limit constraint, if the constraint is violated, the constraint is punished by an fitness function, wherein the fitness function is expressed as: ; Wherein, P H represents the total penalty amount of the height constraint violation, calculates the sum of all stack-bit overrun heights, P o represents the total penalty amount of the warehouse date constraint violation, and omega 1 、ω 2 、ω 3 is a weight coefficient, wherein omega 1 =1,ω 2 =50,ω 3 =30.
8. The method for optimizing stack allocation of steel plates based on LOT combination and split recombination according to claim 1, wherein in step S4, initial temperature T init =100 ℃, cooling rate alpha=0.985 and lowest temperature T min =0 ℃ of a simulated annealing algorithm are set, and the set acceptance probability is expressed as the following formula: ; Wherein T is the current temperature, E is the objective function value of LOT splitting, namely the minimum stack number occupation; let the energy difference Δe denote the difference between the number of occupied stacks corresponding to the new LOT splitting scheme and the number of occupied stacks corresponding to the LOT splitting scheme currently being evaluated, i.e.: ; and when the delta E is more than or equal to 0, accepting the new resolution scheme, and if the delta E is more than or equal to 0, accepting the new resolution scheme by using the probability P, and if the new resolution scheme is not acceptable, replacing the individuals with the largest fitness in the original population with the individuals with the smallest fitness in the new population so as to update the population.
9. The method for optimizing stack allocation of steel plates based on LOT combination and split recombination according to claim 1, wherein in step S5, a population is generated based on initial stack information, simulated annealing parameters are initialized, then an algorithm is circularly executed with the aim of minimizing stack occupation number, the following steps are evaluated, if the population fitness is not optimal, selection, crossover and mutation operations are performed, then the algorithm introduces a simulated annealing process, optimizing screening and acceptance are performed on offspring individuals according to Metropolis criteria, and then cooling is performed, and the iterative process is continued until an optimal solution meeting conditions is output.
10. The optimization method for steel plate stack position allocation based on LOT combination and splitting recombination according to claim 9, wherein when LOT splitting recombination is performed, the encoding mode based on LOT splitting is expressed as follows: ; Wherein k original is the unique identifier of the original LOT, k split_id is the serial number of the sub LOT and is used for splitting different parts detached from the same original LOT, S ksplit is a steel plate set of the sub LOT, h ksplit is the height attribute of the sub LOT, and o ksplit is the date of delivery of the sub LOT; The method comprises the steps of carrying out population initialization by adopting a two-stage repairing method, traversing each stack position in randomly generated individuals in the first stage, carrying out first-round verification and repair, for the stack position which violates the constraint of the delivery date, placing a non-conforming LOT or sub-LOT on the upper layer of the stack position according to the LOT or sub-LOT with the early delivery date, moving out the non-conforming LOT or sub-LOT, for the stack position which violates the high constraint, selecting to move out the LOT which can relieve the overrun problem, temporarily storing all moved out units in a central Pool to be allocated instead of occupying a new stack position immediately, and attempting to redistribute the units in the Pool to the existing stack position which meets the double constraint in the current solution by adopting a first-time adaptive decrementing FFD heuristic, and creating a new stack position for the rest units in the Pool only when the proper stack position cannot be found; The mutation operation comprises splitting mutation, merging mutation and exchange mutation, wherein the mutation operation needs to meet the stack height constraint and the delivery date constraint, otherwise, the mutation operation is regarded as invalid, and the individual can recover the state before mutation.

Description

Steel plate stack position allocation optimization method based on LOT combination and split recombination Technical Field The invention relates to the technical field of steel plate stack position distribution, in particular to a steel plate stack position distribution optimization method based on LOT combination and split recombination. Background The shipyard steel plate yard is a shipyard raw material supply core, bears the key functions of raw material storage, turnover and supply and demand connection, and determines the scheduling rhythm of the subsequent production links. However, in the existing yard management practice, a long-standing core contradiction is that the stacking sequence (warehouse-in sequence) and the processing sequence (warehouse-out sequence) of the steel plates are contradictory, so that the occurrence rate of invalid turnover plates is high. On one hand, under the leading of manual decision, the stack position distribution and the stacking have limitation on the processing of steel plates and storage yard information, and on the other hand, the existing stacking mode does not consider the processing characteristics of the steel plates, so that the steel plates with different characteristics are stacked in an aliasing way. The unordered and low-efficiency current situation not only causes the prolongation of the operation time and the rapid increase of the operation cost, but also becomes the bottleneck for restricting the modern shipbuilding mode to lean and intelligent transformation. Therefore, optimizing the stack allocation of steel sheets is a highly desirable problem for shipyards. Disclosure of Invention Aiming at the problems, the invention aims to provide the steel plate stack position allocation optimization method based on LOT combination and split recombination, which can reduce the number of occupied stack positions and improve the storage yard space utilization rate on the premise of ensuring zero turning plate operation. The technical scheme is that the steel plate stack position allocation optimization method based on LOT combination and splitting recombination comprises the following steps: S1, acquiring parameters of a steel plate, including LOT, dimension specification and delivery date of the steel plate, defining a LOT distribution unit, and determining parameters of a storage yard; s2, designing a steel plate stack position distribution rule and an optimization model; s3, setting genetic algorithm parameters, representing the mapping relation between LOT and the stack position through chromosome coding, and solving to obtain a preliminary stack position distribution scheme through selection, crossing and mutation operations; S4, designing a genetic algorithm for simulating annealing optimization; S5, carrying out LOT splitting and recombining operation at the second stage through a mixing algorithm to obtain a final stack position distribution scheme. Further, in step S1, parameters of the yard are determined by using the LOT, i.e., LOT, of steel plates as core units for stack allocation, including the length x width of each stack in the yard and the capacity of steel plates that can be stacked. Further, in step S2, the stack allocation rule is that different LOTs stacked in the same stack should be placed above the LOT with the earlier date of delivery, while stacking, the optimization model is the minimum number of occupied stacks, and the mathematical model is expressed as: ; Where j= [1,2, ], m ] is the set of available stack bits, y j is 1 if stack bit J is used at least once, otherwise is 0. Optimally, the corresponding constraint conditions of the steel plates in stack position distribution comprise: Each sheet must be assigned a stack, expressed as: ; Wherein i= [1, 2..n ] is a set of steel plates, wherein each steel plate I is uniquely subordinate to one LOT, which represents the kth LOT allocation unit, wherein K e K, k= [1, 2..p ] is a composite LOT, wherein each index corresponds to one unique segment-LOT combination, 1 if the steel plate I is allocated to the stack j, x ij, otherwise 0; each layer of each stack is at most stored with one steel plate expressed as: ; The height of the steel plates stacked in the stack position is not more than the height limit of the stack position, and the steel plates are expressed as follows: ; wherein t i is the thickness of the steel plate i in the subordinate LOTk, and H j is the height of the stack j; The delivery time of the upper steel plate in the stack is earlier than that of the lower steel plate, and is expressed as: ; Wherein, o i is the delivery date of the upper layer steel plate i in the stack position, f ij is the layer number of the upper layer steel plate i in the stack position j, o i' is the delivery date of the lower layer steel plate i 'of the steel plate i in the stack position stack, and f i'j is the layer number of the lower layer steel plate i' in the stack position j; Ensure that If a