CN113343570A

CN113343570A - Dynamic picking method and system considering relevance of picking order

Info

Publication number: CN113343570A
Application number: CN202110640500.1A
Authority: CN
Inventors: 孙军艳; 陈智瑞; 牛亚儒; 闫春妍; 李晶晶
Original assignee: Shaanxi University of Science and Technology
Current assignee: Shaanxi University of Science and Technology
Priority date: 2021-06-08
Filing date: 2021-06-08
Publication date: 2021-09-03

Abstract

The invention discloses a dynamic picking method and a dynamic picking system considering the relevance of picking orders, which belong to the technical field of dynamic picking path optimization of warehouses. The invention can reduce the time for adjusting the goods position of the warehouse with too high frequency of adjusting the goods position, improve the efficiency, simplify the warehousing operation process, reduce the logistics cost, accelerate the goods circulation and improve the economic benefit of enterprises. The method has positive significance for improving the goods picking efficiency and the service level of the whole logistics distribution center, and the conclusion of the model and the numerical simulation can provide decision reference for planning the goods picking path for the e-commerce and enterprise.

Description

Dynamic picking method and system considering relevance of picking order

Technical Field

The invention belongs to the technical field of dynamic picking path optimization of warehouses, and relates to a dynamic picking method and system considering relevance of picking orders.

Background

With the increasing commodity logistics traffic, the picking operation activity is the link of the distribution center with the most intensive labor force and the most cost occupied capital in the warehouse operation. The picking operation optimization mainly comprises order batching, path optimization and the like, wherein the order batching is a premise, the path optimization is a key, and the picking operation optimization can effectively improve the picking efficiency. At present, the warehousing structure can not adapt to the rapid change of commodity demands, huge cost needs to be paid for regularly adjusting and optimizing the warehousing structure, the familiarity degree of a goods picker is greatly influenced, the difficulty of operation field management is increased, and the dynamic goods picking method for picking goods and adjusting the goods position without changing the total storage position of the goods is an effective method for improving the operation efficiency of a distribution center.

Chinese patent CN201911127467.1 discloses a seeding type picking method which mainly comprises three steps: the seeding type goods picking method has the advantages that the efficiency is high, the occupied area of matched equipment is small, the error rate of actual operation is low, the method is flexible, good compatibility is realized, and the actual use value has certain limitation; chinese patent CN201910708249.0 proposes a method and device for processing a picking task to synchronize the task of picking from a persistent database to a file storage database in a cache; analyzing a task list to be picked in the file storage database, and determining a task in the analyzed task list; filtering the tasks in the analyzed task list according to a configurable rule; storing the filtered tasks to a picking queue. The method reduces the pressure of the structured query language for querying the persistent database, so as to flexibly monitor the task to be picked, reduce the change of the task list structure and improve the processing efficiency of the picking task, but the method only optimizes the order before picking and does not provide a specific picking method. Chinese patent CN201811497884.0 proposes a method for implementing a dynamic picking method, which is to complete the picking task of the current picking order, and simultaneously transfer all the goods to be picked corresponding to the goods position to be adjusted in the next picking order of the current picking order to the warehouse entrance and exit, and pick the goods at the warehouse entrance and exit when the picking task of the next picking order of the current picking order is completed, so as to implement the dynamic picking method, thereby effectively improving the picking efficiency of the distribution center, but not considering the problem of order path coincidence in the dynamic picking process.

Disclosure of Invention

The invention aims to overcome the defect that the order and the order can not be associated and processed by the order picking processing method in the prior art, and provides a dynamic order picking method and system considering the association of the order picking.

In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:

a dynamic picking method considering the relevance of picking orders, comprising the following steps:

acquiring a picking order, and sorting the picking order according to the association degree based on the association rule;

adjusting the goods picking orders and goods corresponding to the goods picking vehicles based on the vehicle information of the goods picking vehicles;

setting a picking period, and establishing a picking path optimization model;

and acquiring the shortest picking path based on the picking path optimization model, and picking the picking order and adjusting the goods based on the shortest picking path until the picking of the picking order is finished.

Preferably, the relevancy ranking is specifically that a picking order similar matrix is constructed based on the types, the quantities, the storage locations and the emergency degrees of the commodities in the picking order, and then the relevancy ranking is performed based on the picking order similar matrix.

Preferably, the order and the cargo space corresponding to the truck are adjusted, specifically:

firstly, acquiring the remaining capacity of picking and picking trucks of a plurality of picking orders based on the relevance sorting result, and determining the goods and the quantity to be adjusted; and then, acquiring the residual volume of the picking truck based on the vehicle information of the picking truck, and adjusting the picking order and the goods position based on the residual volume.

Preferably, the establishment process of the picking path optimization model is as follows: setting a picking period, taking the adjusted picking order integration as a target unit in the picking period, and taking the shortest time for completing the picking task as a path optimization target in the picking period to establish a picking path optimization model.

Preferably, the picking path optimization model is specifically:

wherein,

in the above formula, s is cargo; k is a pick-up order; k is a picking order set; i is a goods position set; h is the order of picking order of the picking order k; u is the total picking time of all picking orders; t is t_uThe time required to pick up each unit item; t is_kThe picking time of the picking order k; t'_kThe picking time after the dynamic adjustment of the picking order k is carried out; x is the number of_kiThe number of the ith picking position in the kth picking order; s_kThe distance of the picking path of the kth picking order;

whether the order k +1 contains a goods position i or not is judged; alpha is a common goods position set of the order picking k and the order picking k + 1; beta is a cargo space set to be adjusted; p_kh，e_ks，e_ks'Are all 0,1 variables, P_khRepresents P when the order of picking of the order k is h_khIs 1, otherwise is 0; e.g. of the type_ksRepresenting e when picking the adjusted position s during the picking of the order k_ksTaking 1, otherwise, taking 0; e.g. of the type_ks'Representing e when placing the adjusted position s during the picking of the pick slip k_ks'Take 1, otherwise 0.

Preferably, the shortest pick path is solved using a genetic simulated annealing algorithm.

Preferably, the specific process of the genetic simulated annealing algorithm is as follows:

coding a goods location set to be adjusted;

constructing initial individuals in a solution space to form an initial population, and establishing an initial population fitness function;

calculating individual fitness values based on the initial population fitness function, and performing selection, crossing and mutation operations to record excellent individuals;

and (4) annealing operation is carried out by taking the good individuals as the initial solution of the simulated annealing operator, and the fitness values of the parent generation and the offspring are calculated.

Preferably, the picking position to be adjusted is coded, and the mathematical model of the dynamic picking path optimization problem is solved by adopting a genetic simulation annealing algorithm to obtain an optimal picking path scheme. An integer coding mode is adopted in the hybrid genetic simulated annealing algorithm, and parameters are set as follows: number of iterations T2000, population size N80, initial temperature T₀50000, annealing coefficient a 0.99, variation probability p_m0.4, cross probability p_c＝0.8

Preferably, picking order and adjusting goods position are carried out based on the shortest picking path, and the specific process is as follows:

optimizing the shortest picking path of the picking order k, and putting the goods to be adjusted on a shelf and placing the goods at a specified position while finishing the operation of the picking order k;

optimally solving the shortest picking path of the picking order k +1 after the goods position is adjusted, and updating k +1 into k;

judging whether the order k is the last order,

if yes, finishing the sorting of the sorting order k and extracting the goods, and finishing the sorting of all orders;

if not, the relevance between the order picking sheet k and the remaining order picking sheets is solved, and the relevance is maximized.

A dynamic picking system that considers picking order relevance, comprising:

the information acquisition module is used for acquiring the order picking list and sorting the order picking list according to the association rule; simultaneously acquiring vehicle information of the goods picking vehicle, and adjusting the goods picking order and goods corresponding to the goods picking vehicle;

the model establishing module is interacted with the information acquisition module and used for establishing a goods picking path optimization model based on the adjusted goods picking orders and goods;

and the calculation processing module is interacted with the model establishing module and is used for acquiring the shortest picking path based on the picking path optimization model, and picking orders and adjusting goods based on the shortest picking path until the picking of the picking orders is finished.

Compared with the prior art, the invention has the following beneficial effects:

the invention discloses a dynamic picking method considering relevance of picking orders. Aiming at the problems that in the warehouse-out operation process, the step optimization is carried out, the warehouse structure cannot adapt to the rapid change of commodity requirements, huge cost needs to be paid for regular adjustment and optimization of the warehouse structure and the like, the invention provides the idea of 'cargo position adjustment and cargo picking path' collaborative optimization, the cargo picking process and the cargo position adjustment process are combined into one, the association degree information among the cargo picking orders is mined by using the association rules and used as the input information of the cargo position adjustment and the path optimization, the problem of the cargo position dynamic adjustment and the path joint optimization is solved, and the integral optimization of the cargo position adjustment and the path optimization is realized. The method can perform goods position adjustment operation while picking the goods, can reduce the time for adjusting the goods position for the warehouse with high frequency of adjusting the goods position so as to improve the efficiency, simplify the warehousing operation process, reduce the logistics cost, accelerate the goods circulation and improve the economic benefit of enterprises. The method has positive significance for improving the goods picking efficiency and the service level of the whole logistics distribution center, and the conclusion of the model and the numerical simulation can provide decision reference for planning the goods picking path of the e-commerce enterprise.

Furthermore, similarity among orders is mined based on information such as commodity types, quantity, storage positions and emergency degree in the picking order, a single-phase similarity matrix for picking is constructed, and then relevance ranking is carried out based on the single-phase similarity matrix for picking as the basis for sorting the picking order.

Furthermore, a picking period is set, the adjusted picking order integration is used as a target unit in the picking period, the shortest time for completing the picking task is used as a path optimization target in the picking period, and a picking path optimization model is established. And solving by using a genetic simulated annealing algorithm to obtain the shortest picking path.

The invention discloses a dynamic picking system considering the relevance of picking orders, which is based on the collaborative optimization thought of 'picking and simultaneously adjusting the goods location', simultaneously considers the relevance among the picking orders, utilizes a relevance rule to mine the relevance information among the picking orders, designs a collaborative optimization dynamic picking method of sorting orders of the picking orders, adjusting the goods location and optimizing the path, utilizes the relevance rule to sort the picking orders, establishes a mathematical model taking the shortest total picking time as a target, and adopts a genetic simulation annealing algorithm to solve the model. Under the background of explosive increase of goods picking tasks in the e-commerce industry, the system can comprehensively and systematically optimize the goods picking operation and improve the goods picking efficiency.

Drawings

FIG. 1 is a flow chart of dynamic pick path optimization according to the present invention;

FIG. 2 is a flow chart of the order adjustment phase of the present invention;

FIG. 3 is a flow chart of a hybrid genetic simulated annealing algorithm of the present invention;

FIG. 4 is a layout of a dual-zone warehouse of the present invention;

FIG. 5 is a diagram of a path for adjusting the dynamic picking method of the picking order 3 in consideration of order relevance according to the present invention;

FIG. 6 is a diagram of a path for adjusting the dynamic picking method of the picking order 2 in consideration of order relevance according to the present invention;

FIG. 7 is a diagram of a path for adjusting the dynamic picking method of the picking order 5 in consideration of order relevance according to the present invention;

FIG. 8 is a path diagram of the dynamic picking method adjustment of the picking order 1 in consideration of order relevance according to the present invention;

FIG. 9 is a diagram of an adjustment path of the dynamic pick method for picking orders 4 according to the present invention with consideration of order relevance.

Detailed Description

The invention is described in further detail below with reference to the accompanying drawings:

example 1

step 1) acquiring a picking order, and sorting the picking order according to the relevance degree based on the relevance rule;

step 2) adjusting the goods picking orders and goods corresponding to the goods picking vehicles based on the vehicle information of the goods picking vehicles;

step 3) setting a picking period and establishing a picking path optimization model;

and 4) acquiring the shortest picking path based on the picking path optimization model, and picking orders and adjusting goods based on the shortest picking path until the picking of the picking orders is finished.

Example 2

A dynamic order picking method and system considering order picking relevance specifically comprises the following steps:

step 1): firstly, all the picking orders in the t period are obtained, the picking orders are sorted according to the association rule,

the picking order in the same period mainly comprises various information such as commodity types, quantity, storage positions, emergency degree and the like, and the picking order correlation matrix is constructed mainly by considering item similarity and the picking order emergency degree. The method comprises the following specific steps:

(101) the order collection received in the same time period is X_n＝{x₁,x₂Λ,x_n}。

First extracting X_nItem in (1) and picking urgency information. Because different index data magnitudes are different, the item similarity is mainly based on the correlation coefficient, and the order urgency degree is based on the order predicted delivery time; therefore, the order similarity and the estimated delivery time of the two indexes are calculated and normalized respectively, and then the importance weight omega of the two indexes is given based on expert evaluation₁，ω₂Finally, the total correlation value is the weight of each indexAnd (c). And obtaining a final relevance value.

(102) Calculating the similarity matrix L (L) of the items in the order collection_i,j＝l_j,i)

Based on the proportion of common item categories in the two picklists, a measure of how similar the picklists i and j are is defined as follows:

(103) to l_ijCarrying out normalization treatment to obtain normalized l'_ijConverting the non-diagonal elements in the matrix L by using the following formula to obtain a new processed matrix L'

Wherein l'_ijMin l elements of the new matrix after normalization_i,jAnd max l_i,jWhich represent the minimum and maximum values, respectively, of the elements on the off-diagonal of the matrix L.

(104) Calculating the urgency matrix T (T) of the order collection_i,j＝t_j,i)

Defining; the pick-up urgency formula between two pick-up orders is as follows

Wherein, T_i ^αEstimated delivery time, T, for representative order i_i ⁰Representative orderActual start of processing time for a single i, T_j ^αEstimated delivery time, T, for representative order j_j ⁰Representing the actual start processing time of order j, the pick urgency matrix T between all pick orders is:

(105)

the matrix L' and the matrix T are fused to obtain a total correlation matrix R

Wherein r is_i,jIs an element in the total correlation matrix, r_i,j＝w₁l′_i,j+w₂t_i,j；w₁And w₂Is the index importance weight.

(106) And selecting the maximum value in the total similarity matrix, extracting two picking orders corresponding to the maximum value, wherein the picking order is k and k +1, and the initial state k is 1.

Step 2): according to the relevance sorting result, the sorting order of the n sorting orders and the residual capacity of each sorting vehicle for the sorting orders are obtained, and the goods and the quantity to be adjusted are determined, so that the residual capacity of the sorting vehicles is utilized to the maximum extent, and the total sorting time and distance are shortened. And adjusting goods on the goods position to be adjusted in the pick order with the sequence of k +1 while finishing the pick order task with the sequence of k, and placing the adjusted goods on the vacant position arranged on the door. The method comprises the following specific steps:

(201) extracting two picking orders with the maximum total relevance degree R, extracting information of the two picking orders, wherein the sequence of the information is k, k +1, the initial state k is 1, and the goods position set of all goods contained in the information is A_i；

(202) Determining that the same goods position set of the pick order with the rank k and the pick order with the rank k +1 is alpha; the number of common goods is calculated using equation (4): formula (5) calculates the common cargo volume

In the formula, i is a goods picking position in the goods picking order; q is the common cargo volume in the two picking orders; n is_iThe total goods quantity in the two picking orders is obtained;

whether a picking path of the picking order with the sequence of k contains a goods position i or not;

whether the pick order with the sequence of k +1 contains a goods position i or not; q. q.s_iThe volume of goods (the number of the goods is more than or equal to 1, and the specific number is determined by the remaining volume of the picking truck) at the picking position i.

(203) The remaining capacity of each goods picking order and goods picking truck is f, and the total volume of the goods needs to be adjusted to be less than or equal to the remaining capacity of the goods picking truck. Namely, it is

When the path of the k-ordered pickups and the k + 1-ordered pickups have the same goods,

and

all take the value of 1.

Wherein f is the remaining volume of the goods picking vehicle; i is a goods picking position in the goods picking order; q is the common cargo volume in the two picking orders;

picking path for picking orders ordered as kWhether the goods position i is included;

(204) A set of adjustable cargo spaces beta is determined. The specific flow chart is shown in fig. 2.

Wherein: f is the remaining volume of the goods picking vehicle; i is a goods picking position in the goods picking order; q is the common cargo volume in the two picking orders; n is_iThe total goods quantity in the two picking orders is obtained;

whether the pick order with the sequence of k +1 contains a goods position i or not; q. q.s_iThe volume of goods (the number of the goods is more than or equal to 1, and the specific number is determined by the remaining volume of the picking truck) at a picking position i; alpha is the collection of the common goods positions of the order sorted by k and the order sorted by k +1, and beta is the collection of the goods positions to be adjusted.

Step 3): the objective of dynamic picking path optimization is to find a picking mode which minimizes the time taken to complete the picking task and the picking path, and a specific mathematical model is expressed as follows.

Some parameters are first defined as follows:

the model assumes that:

picking the goods according to a picking order, wherein the idle waiting time is ignored;

the order picker can walk along 2 directions of the longitudinal roadway and the transverse channel;

the distribution of the goods picking area mainly comprises: each cargo space has the same size and the same bearing capacity, and only one cargo is stored in each cargo space;

the goods shortage phenomenon does not exist in the goods picking process, namely the goods supplementing time is ignored;

the size of each order does not exceed the maximum picking amount of a picker;

the cargo space is large enough to ensure that each SKU can be stored.

The corresponding parameters and variables of the model are defined as follows:

k: a collection of pick-up orders, K belongs to K,

i: a set of cargo space, I ∈ I,

h: the order of picking the goods in the order of picking the goods,

u: the total picking time of all the picking orders,

t_u: the time required to pick up each unit item,

T_k: the picking time of the picking order k,

T′_k: the picking time after the dynamic adjustment of the picking order k,

x_ki: the number of orders in the ith pick order in the kth pick order,

S_k: the distance of the k-th order picking path,

whether the pick order k +1 contains the cargo space i,

α: the picking order k and the picking order k +1 share a goods space set,

beta: the set of the cargo space to be adjusted,

P_kh: whether the picking order of the picking order k is h,

e_ks: the adjusted goods position s is picked in the picking process of the picking order k,

e'_ks: the goods location s which is adjusted is placed in the picking process of the picking order k.

An objective function:

wherein,

constraint conditions are as follows:

equation (6) represents the total pick time for all picklists; the formula (7) represents the time of the original goods picking path before the goods location is adjusted; the formula (8) represents the total time of the original goods picking path and the adjusted goods position; equation (9) represents the time of the replanned path after the adjusted cargo space is subtracted from the pick order sorted as k + 1; equation (10) indicates that the order of each pick order is unique; equation (11) indicates that only one picking order is assigned per picking order; the formula (12) shows that the picking order with the sequence k is monotonously adjusted, and the number of goods positions is multiple; the formula (13) shows that the quantity of the goods to be adjusted picked and the quantity of the goods to be adjusted placed in the order of k are equal; if the sorting order of the picking order k is h, the value is 1, otherwise the value is 0; if the sorting order k in the formula (15) is used for sorting the goods to be adjusted in the sorting process, the value of the goods position s to be adjusted is 1, otherwise, the value is 0; and (3) if the goods to be adjusted are placed in the picking order with the sequence of k in the picking process of the formula (16), taking the value of 1, otherwise, taking the value of 0.

In the formulas (6) to (16), f is the remaining volume of the picking truck, i is the picking position in the picking order, Q is the total goods volume between the front and rear picking orders, and n_iIn order to share the quantity of goods among the picking orders,

the picking path for the order k includes a cargo space i,

whether the pick order with the order of k +1 contains the goods position i, q_iThe volume of goods at the picking position i, alpha is the picking order with the order of k and the picking order with the order of k +1 share the picking position set, beta is the picking position set to be adjusted, P_hkWhether the order of picking is k, x or not_kiNumber of items of ith pick-up space of kth pick-up order, e_ksFor picking adjusted goods space s, e 'in picking process of picking order k'_ksFor placing adjusted goods space S, S in the picking process of picking order k_kDistance, t, of the picking path for the kth order_uTo pick the time required per unit item, U is the total pick time for all picklists.

The empty goods shelves are arranged beside the warehouse entrance, goods on the goods position to be adjusted are moved to the empty goods shelves from the existing storage positions according to the required quantity during picking, and the goods on the picking order are taken from the arranged goods shelves and the goods on the goods position to be adjusted on the next picking order are placed when the next picking is finished. Obtaining all picking orders in a t period, sequencing the picking orders according to the relevance degree of the relevance rules, entering a goods position dynamic adjustment stage, determining the next picking order and goods to be adjusted by considering factors such as the volume of a picking truck, task balance and the like, solving the shortest path of the picking orders sequenced into a k picking order and k +1 picking orders, and putting the goods to be adjusted sequenced into the k +1 picking order on a set empty goods shelf when a picker finishes the operation of sequencing into the k picking order, namely finishing the goods position adjustment while picking the goods; and optimally solving the shortest picking path of the order of the adjusted goods into a picking order, and updating k +1 into k. Judging whether k is the last picking order, if so, finishing picking of the picking orders sorted for k and taking away goods on the goods position, and finishing picking of all orders; if not, the relevance between the order picking k and the remaining order picking is solved to maximize the relevance, and the steps are repeated to complete the picking tasks of all the order picking, and the specific steps are shown in fig. 1.

Step 4): and (4) carrying out simulation analysis on the total time objective function model by using MATLAB based on a hybrid genetic simulation annealing algorithm to obtain the optimal dynamic picking sequence and the shortest picking time. Referring to fig. 3, a population is initialized by using a genetic algorithm, the genetic algorithm is easy to fall into a local optimal solution in an iterative process, the local optimal solution is skipped from the search to obtain a global optimal solution by using the random search probability of the simulated annealing algorithm, and the detailed steps based on the hybrid genetic simulated annealing algorithm are as follows:

(401) coding design, in this context using natural number coding, chromosome C ═ i₁,i₂,i₃Λ i_n]Wherein the element (gene) [ i_n]Is [1, n ]](n is the number of picking positions), wherein the gene number corresponds to the sorting of the picking positions, and the gene value is the serial number of the picking positions needing to be picked.

(402) Initializing a population M₀Randomly generating N initial chromosomes C_jPopulation (j is more than or equal to 1 and less than or equal to N), namely randomly generating N order picking sequences, wherein the length of the gene sequence of the chromosome depends on the actual condition of picking orders.

(403) A fitness function f, wherein the fitness function represents the reciprocal of the total distance of the picking order, and the calculation formula is as follows:

(404) selecting operation, first making the cycle count variable gen equal to 0 and the temperature T_nThe selection is performed by roulette, the angle of the sector is proportional to the fitness of the individual, and the higher the fitness, the greater the probability of selection.

(405) Performing crossover operation by first performing crossover mutation by using partial matching crossover method with crossover probability P_cAnd judging whether the individuals carry out the cross operation or not.

(406) Mutation operation by single point mutation method with mutation probability P_mAnd new populations are generated.

(407) Simulating annealing operation, calculating parent C_jAnd progeny C'_jOf (d), i.e. f (C)_j) And f (C'_j) And theta is [0, 1 ]]Constant between, then the temperature calculation formula is T_n+1＝θ×T_nDuring the simulated annealing, the new state is accepted with a certain probability given by the Metropolis criteria.

(408) And judging that the loop counting variable gen is less than Maxgen, if gen is equal to gen +1, returning to the step (404), and otherwise, turning to the step (49).

(409) If T_n<T_endIf the algorithm is finished, returning to the optimal solution; otherwise, executing a cooling operation T_n+1＝θ×T_nAnd returning to the step (403).

Example 3

The following description will be made by taking a two-zone type warehouse as an example. Suppose that the double-area warehouse is composed of 3 parallel transverse channels and 10 longitudinal channels, the width of each channel and the width of the lane are equal, the shelves in the first column and the last column of the longitudinal lane are single-row shelves, and the rest shelves are back-to-back shelves, as shown in fig. 4.

The width of each roadway is 1m, the width of each transverse channel is 2m, the number of shelves in each row between the two transverse channels is 20, and the length and the width of each shelf are both 1 m. The number of the storage positions in the warehouse is 800, and the serial number of each storage position is unique. The entrance and exit of the warehouse are arranged at the lower left corner, the order picker can walk along the south and north directions, and the starting point and the ending point of the order picker are both the entrance and exit of the warehouse. Each channel and each lane are equal in width, and the number [ a, b, c, v ] of each storage position is equal to that of the lane where the goods position is located, wherein a is 1, 2, 3.. 10; b represents that if the goods are taken on the left goods shelf, b is 0, and if the goods are taken on the right goods shelf, b is 1; c represents the serial numbers of the goods positions in the same row of goods shelves, and the serial numbers are 1, 2, 3.. 40 from bottom to top; v represents the unit volume of the goods on each depot. For example, the number of the storage positions is [4, 0, 6, 2] to indicate that the picked goods are in the 6 th storage position from bottom to top on the 4 th roadway, the left side shelf and the unit volume is 2.

Distance d between picking points_ijThe specific calculation method comprises the following steps:

assume that any two pick points i, j are labeled [ a ]_i，b_i，c_i，v_i]And [ a ]_j，b_j，c_j，v_j]Distance d between the two pick-up points_ijThe following were used:

if two goods sites are in the same picking lane, on the same side of the middle channel, d_ijFor difference between goods numbering

|c_i-c_j|；(a_i＝a_j,b_i＝b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (19)

|c_i-c_j|+1；(a_i＝a_j，b_i≠b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (20)

If the two goods sites are in the same picking roadway and are dispersed at the two sides of the middle channel,d_ijfor difference of goods position number plus width of middle channel

|c_i-c_j|；(a_i＝a_j,b_i＝b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (21)

|c_i-c_j|+1；(a_i＝a_j,b_i≠b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (22)

If two goods sites are in different picking lanes, on the same side of the middle channel, d_ijThe shortest path after the intersection direction of the roadway and the channel is selected, namely the transverse distance between the minimum goods position number difference value and the goods positions.

min(c_i+c_j,40-(c_i+c_j))+|a_i-a_j|×3,(a_i≠a_j,b_i＝b_j,0＜c_i,c_j≤20) (23)

min(c_i+c_j,40-(c_i+c_j))+|a_i-a_j|×3+1(a_i≠a_j,b_i＜b_j,0＜c_i,c_j≤20) (24)

min(c_i+c_j,40-(c_i+c_j))+|a_i-a_j|×3-1,(a_i≠a_j,b_i＝b_j,20＜c_i,c_j≤20) (25)

min(c_i+c_j-40,80-(c_i+c_j))+|a_i-a_j|×3,(a_i≠a_j,b_i＜b_j,20＜c_i，c_j≤40) (26)

min(c_i+c_j-40,80-(c_i+c_j))+|a_i-a_j|×3+1,(a_i≠a_j,b_i＜b_j,20＜c_i,c_j≤40) (27)

min(c_i+c_j-40,80-(c_i+c_j))+|a_i-a_j|×3-1,(a_i≠a_j,b_i＞b_j,20＜c_i,c_j≤40) (28)

If two goods sites are in different picking lanes, distributed on both sides of the central channel, d_ijFor the difference between the goods position serial numbers plus the width of the middle channel 2m plus the distance between two goods position transverse roadways

|c_i-c_j|+2+|a_i-a_j|×3,(a_i≠a_j,b_i＝b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (29)

|c_i-c_j|+2+|a_i-a_j|×3+1,(a_i≠a_j,b_i＜b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (30)

|c_i-c_j|+2+|a_i-a_j|×3-1,(a_i≠a_j,b_i＞b_j,0＜c_i≤20＜c_jLess than or equal to 40, or 0 less than c_j≤20＜c_i≤40) (31)

Distance between cargo site i and entrance/exit

c_i+|a_i-a₁|×3,(0＜c_i≤20) (32)

c_i+2+|a_i-1|×3,(20＜c_i≤40) (33)

The general parameter settings for the warehouse are detailed in table 1.

TABLE 1 general parameters settings for warehouse

Parameter(s)	Description and setting of parameters
		v	Average walking speed of goods picking personnel is 1m/s
d	Each lane has a width of 1m
		l	Each transverse channel has a width of 2m
γ	The length and width of the goods shelf are both 1m
		t_r	The time required for the picking personnel to pick single goods is 2s
Q	The capacity Q of the picking truck is 80

Firstly, taking 5 picking orders as an example, solving a dynamic picking method model considering the relevance of the picking orders by adopting a genetic annealing algorithm, taking the picking order 3 of fig. 5 as an example, marking a circle as an optimal route of all goods in the picking order 3, marking a square as a coincidence of the picking order 3 and the picking order 2, dynamically adjusting the coincidence to 4 SKUS at an entrance/exit vacant shelf when the task of the picking order 3 is completed, wherein the SKUS comprises

goods positions

251, 297, 449 and 450, re-planning paths of the remaining goods of the picking order 2, namely 6 SKUS which are marked as a circular mark part in fig. 4, marked as a coincidence of the picking order 2 and need to be adjusted, comprise the goods positions 34, 61, 64, 105, 166 and 665, and the like, knowing the goods and the quantity of other picking orders which need to be adjusted, and the increase and decrease of the picking paths and the picking time.

The method provided by the invention is optimized with a dynamic order picking method which does not consider the relevance of order picking orders and a traditional static method for picking orders, wherein the traditional method refers to that the orders are sequentially picked according to the order picking orders, and each order picking process finishes the SKU of the whole order picking order; the non-sorting method refers to a dynamic sorting method which does not carry out the digging and sorting of the relevance degree of the goods space on the sorting order.

The dynamic picking method considering the relevance of the picking orders has the picking order of 3-2-1-4-5, the specific path of each picking order is shown in figures 5-9, the total picking time is 1442 seconds, and the total picking path is 1106 meters. The dynamic picking method without considering the relevance of the picking order has the picking order of 1-2-3-4-5, the total picking time of 1571 second and the total picking path of 1180 meters. The traditional static method has the order of picking the order of 1-2-3-4-5, the total picking time is 1627 seconds, and the total picking path is 1291 m.

The comparative results are shown in Table 2.

TABLE 2 comparative analysis of three picking methods

As can be seen from table 2, under the condition of the genetic annealing algorithm, compared with the conventional picking method, the dynamic picking method after sorting the picking orders shortens the picking time by 8.69% and 11.37%, and shortens the picking distance by 10.86% and 14.33%; compared with the dynamic sorting method without sorting, the sorting time is respectively shortened by 5.9% and 7.38%, and the sorting distance is shortened by 7.37% and 9.3%.

Secondly, in order to verify the superiority of the invention to the picking orders of different scales, simulation experiments are carried out by taking 5, 13 and 54 picking orders as examples. Table 3 shows the comparison results of the solution optimization performed by the genetic algorithm and the hybrid genetic simulated annealing algorithm for different picking methods without picking order scale.

TABLE 3 comparison of picking times(s) for different picking order scales

From table 3, the following conclusions can be drawn:

conclusion 1) for the same batch of picking orders of different scales, the results of the hybrid genetic algorithm are superior to the genetic algorithm no matter what picking method.

Conclusion 2) under the condition of the same algorithm, the sorted dynamic order picking method is better than the non-sorted dynamic order picking result, and compared with the traditional order picking method, the time saving of the sorted dynamic order picking method is more than 3 times that of the non-sorted dynamic order picking method.

Conclusion 3) as the amount of picking orders increases, the more time is saved in the dynamic picking method after sorting by considering relevance of the picking orders, the more remarkable the picking efficiency of the picking mode is improved, and the picking time reduction percentage is increased from 10.03% to 18.38%.

Conclusion 4) the dynamic picking method considering the relevance of the picking order, which is provided by the invention, further optimizes the picking path in the whole period, further shortens the picking path and time, improves the picking efficiency, solves the problem of goods position adjustment, further proves the feasibility of the invention, and provides a basis for the actual warehouse picking operation.

In the verification of the embodiment, the traditional picking method is obtained by adopting the genetic algorithm and the genetic simulated annealing algorithm, the time and the path length under the picking mode are obtained by the method after the relevance of the picking order is not considered and the relevance is considered, the larger the time saving ratio of the picking method is when the picking order quantity is increased is obtained by analyzing the experimental examples with different scales, the more obvious the effect is, and the efficiency of picking personnel can be obviously improved by the dynamic picking mode considering the relevance of the picking order; under the same conditions, the optimization result of the genetic simulated annealing algorithm is better than that of the genetic algorithm.

Example 4

A dynamic picking system that considers picking order relevance, comprising:

In summary, the invention provides a dynamic picking method considering relevance of picking orders and an implementation method thereof, based on a collaborative optimization idea of 'picking and simultaneously adjusting goods location', applies a collaborative optimization dynamic picking method based on association rule mining association degree information among the picking orders, designs order sorting-goods location adjustment-path optimization of the picking orders, sorts the picking orders by using the association rule, establishes a mathematical model with the shortest total picking time as a target, and solves the model by adopting genetic simulation annealing to obtain the shortest picking path.

The dynamic adjustment method optimizes the path of the subsequent order picking orders to the maximum extent without increasing the picking path of the current order picking orders, and can avoid secondary sorting. Meanwhile, the method is not only suitable for picking goods manually, but also suitable for picking electronic tags. In addition, since the adjustment only involves a certain number of goods of the next pick order, there is no adjustment of the whole goods space, so that the existing WMS of the warehouse is not affected. Therefore, the designed dynamic order picking method considering the relevance of the order picking list is simple and convenient, is easy to implement, and further improves the order picking efficiency. The logistics distribution center has positive significance for improving the picking efficiency and the service level of the whole logistics distribution center, and can provide decision reference for planning picking paths for e-commerce enterprises.

The above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.

Claims

1. A dynamic picking method considering relevance of picking orders is characterized by comprising the following steps:

setting a picking period, and establishing a picking path optimization model;

2. The dynamic picking method considering the relevance of picking orders as claimed in claim 1, wherein the relevance ranking is implemented by first constructing a picking order similarity matrix based on the type, quantity, storage location and urgency of the goods in the picking order, and then performing relevance ranking based on the picking order similarity matrix.

3. The dynamic picking method considering the relevance of picking orders as claimed in claim 1, wherein the picking orders and the locations corresponding to the picking trucks are adjusted, specifically:

4. The dynamic picking method considering relevance of picking orders according to claim 1, characterized in that the picking path optimization model is established by the following procedures: setting a picking period, taking the adjusted picking order integration as a target unit in the picking period, and taking the shortest time for completing the picking task as a path optimization target in the picking period to establish a picking path optimization model.

5. The dynamic picking method considering picking order relevance according to claim 1, wherein the picking path optimization model is specifically:

wherein,

6. The method of claim 1 wherein the shortest picking path is solved using a genetic simulated annealing algorithm.

7. The dynamic picking method considering picking order relevance according to claim 6, characterized in that the specific process of the genetic simulation annealing algorithm is as follows:

coding a goods location set to be adjusted;

8. The dynamic picking method considering the relevance of picking orders as claimed in claim 7, wherein the picking positions to be adjusted are encoded, and a mathematical model of a dynamic picking path optimization problem is solved by adopting a genetic simulation annealing algorithm to obtain an optimal picking path scheme;

an integer coding mode is adopted in the hybrid genetic simulated annealing algorithm, and parameters are set as follows: number of iterations T2000, population size N80, initial temperature T₀50000, annealing coefficient a 0.99, variation probability p_m0.4, cross probability p_c＝0.8。

9. The dynamic picking method considering the relevance of picking orders as claimed in claim 1, wherein the picking orders and the goods level adjustment are performed based on the shortest picking path by the following steps:

judging whether the order k is the last order,

10. A dynamic order picking system that considers order picking relevance, comprising: