CN102214321A - Method and system for optimizing loading layout of three-dimensional containers - Google Patents

Abstract

The invention discloses a method and a system for optimizing loading layout of three-dimensional containers, wherein the method comprises finding a group of rough feasible solutions in a solution space by utilizing the random rapid searching capacity, potential parallelism and global convergence of a genetic algorithm, then using the group of rough feasible solutions as the inputs of an ant colony algorithm, and obtaining the optimal loading solution by the positive feedback mechanism, parallelism and capacity of searching better solutions of the ant colony algorithm, thereby realizing the fusion of the genetic algorithm and the ant colony algorithm in the solution to the loading layout problem of containers; the method avoids the defects of solving the loading layout problem of three-dimensional containers by a single algorithm at present, gives consideration to a plurality of important constraint conditions influencing the loading efficiency while considering global search capability, and has a good applicability.

Description

A kind of three-dimensional container loading layout optimization method and system

Technical field

The invention belongs to container loading layout optimization design field, relate in particular to a kind of method and system of three-dimensional container loading layout optimization.

Background technology

Three-dimensional container loading layout optimization method refers under certain constraint condition, during a batch of goods packed into the container according to suitable stowage, so that the capacity utilization of container or loading mass utilization factor maximum, thereby realize container is carried out rationally the effectively method of use, its objective function can be expressed as: $\max Z=\mathrm{\λ}(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{l}_i\·w_i\·h_i\·{\mathrm{\δ}}_i\·m)/V+(1-\mathrm{\λ})\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{g}_i\·{\mathrm{\δ}}_i\·m/G,$ Wherein, li, wi, hi, gi, m represent length, quality, the number of packages of i class goods respectively; V, G represent maximum load volume, the maximum laden mass of container respectively; λ is the 0-1 variable, λ=1 when pursuing one's goal to the capacity utilization maximum, λ=0 when pursuing one's goal to loading mass utilization factor maximum; I is the 0-1 variable, if goods i loads then δ i=1, otherwise δ i=0.

The three-dimensional container loading layout optimization method that prior art provides is many based on single intelligent optimization algorithm, as based on the three-dimensional container loading layout optimization method of genetic algorithm, based on the three-dimensional container loading layout optimization method of ant group algorithm with based on the three-dimensional container loading layout optimization method of heuritic approach etc.Because the algorithm that adopts is single, when solving three-dimensional container loading layout optimization problem, can't all show certain defective in conjunction with the superiority of algorithms of different.For example, based on the three-dimensional container loading layout optimization method of heuritic approach, layout piece bleed strategy, the efficiency of loading that can implement to determine be higher, but it has only been considered the constraint of vanning volume and has considered not enough to the constraint condition of vanning others; Three-dimensional container loading layout optimization method based on genetic algorithm, genetic algorithm as its theoretical foundation has the group ability of searching optimum, extensibility is strong, easily with advantages such as other technologies combine, but owing to do not make full use of system feedback information, feasible search has blindness, often forms redundant iteration when algorithm is found the solution certain limit, and the efficient that causes seeking optimum solution reduces; Three-dimensional container loading layout optimization method based on ant group algorithm, ant group algorithm as its theoretical foundation is a kind of algorithm that combines Distributed Calculation, positive feedback mechanism and the search of greedy formula, has the more excellent ability of separating of very strong search, its renewal by pheromones efficiently converges to optimum solution, but because initial stage pheromones scarcity, it is longer to cause searching for the time that initial stage accumulating information element takies.

Summary of the invention

The purpose of the embodiment of the invention is to provide a kind of method of three-dimensional container loading layout optimization, the three-dimensional container loading layout optimization method that provides with the solution prior art is many based on single intelligent optimization algorithm, when solving three-dimensional container loading layout optimization problem, can't all show the problem of certain defective in conjunction with the superiority of algorithms of different.

The embodiment of the invention is achieved in that a kind of method of three-dimensional container loading layout optimization, said method comprising the steps of:

Treat the vanning goods and encode, generate and wait a plurality of initial cargo collection of goods of casing, and utilize genetic algorithm to generate the filial generation goods collection of described initial cargo collection;

Calculate filial generation each individual fitness value of cargo consolidation and each individual fitness value of parent cargo consolidation thereof, when relative each the individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation no longer increases, corresponding filial generation goods collection is carried out the individuality decoding, obtain feasible solution;

The set that a plurality of feasible solutions that a plurality of initial cargo set pairs are answered constitute is as the input of ant group algorithm, utilizes the ant group algorithm iterative search to obtain the optimum vanning scheme of the described goods of waiting to case.

Wherein, the described step of utilizing genetic algorithm to generate the filial generation goods collection of described initial cargo collection can also may further comprise the steps:

Initialization genetic algorithm controlled variable;

Calculate described initial cargo according to fitness function and concentrate each individual fitness value;

According to described fitness value that calculates and the variation probability that prestores, be the basis iterative processing of selecting, intersect, make a variation with described initial cargo collection, obtain a plurality of continuous filial generation goods collection of described initial cargo collection.

Further, described fitness function can satisfy relational expression:

F＝(l _i·w _i·h _i/L _j·W _j·H _j)×100％

Wherein, F is a fitness function; l _iBe the length of i class goods, w _iBe the width of i class goods, h _iIt is the height of i class goods; L _jBe the length of container, W _jBe the width of container, H _jHeight for container.

Further, described selection is handled can adopt optimum conversation strategy and roulette back-and-forth method, at this moment, described fitness value that described basis calculates and the variation probability that prestores, with described initial cargo collection is the basis iterative processing of selecting, intersect, make a variation, and the step that obtains a plurality of continuous filial generation goods collection of described initial cargo collection can also may further comprise the steps:

According to the ideal adaptation degree value that calculates and optimum conversation strategy and roulette back-and-forth method, determine the selection probability of concentrated each individuality of described initial cargo;

Selection probability according to described each individuality of determining is selected two father's individualities in described initial cargo collection;

According to the variation probability that prestores described two father's individualities of selecting are made a variation and to handle or cross processing, two father's individualities after handling are inserted into next continuous filial generation cargo consolidation of described initial cargo collection, and calculate next each individual fitness value of continuous filial generation cargo consolidation of described initial cargo collection;

According to each the individual fitness value of current filial generation cargo consolidation and optimum conversation strategy and the roulette back-and-forth method that calculate, determine the selection probability of current each individuality of filial generation cargo consolidation;

Selection probability according to current each individuality of filial generation cargo consolidation of determining is selected two father's individualities in current filial generation goods collection;

Handle or cross processing according to described two father's individualities of selecting being made a variation according to the variation probability that prestores, two father's individualities after handling are inserted into next continuous filial generation cargo consolidation of current filial generation goods collection, thereby iteration obtains a plurality of continuous filial generation goods collection of described initial cargo collection.

Wherein, described corresponding filial generation goods collection is carried out individuality decoding, the step that obtains feasible solution can also may further comprise the steps:

Take out a goods to be cased in turn from corresponding filial generation cargo consolidation, calculate take out described wait the to case volume of goods according to the described goods information of waiting to case that takes out;

According to vanning constraint condition V '+l _iW _iH _iDuring≤V judges whether the described goods of waiting to case that takes out can pack into the container, wherein, the measurement of cargo of V ' for having packed into the container, l _iW _iH _iBe the volume of the described goods of waiting to case that takes out, V is the useful volume of container;

According to judged result, when the goods described to be cased that judge to take out can pack into the container when middle, the volume of the described goods of waiting to case that takes out is added among the described measurement of cargo V ' that has packed into the container, when the goods described to be cased that judge to take out cannot pack into the container when middle, take out next goods to be cased in turn from corresponding filial generation cargo consolidation, case volume and the measurement of cargo sum that has packed into the container of goods during greater than the useful volume of described container when corresponding filial generation cargo consolidation goods all to be cased waiting of having got or taken out, obtain the feasible solution after corresponding filial generation goods collection transforms.

Wherein, the set that the described a plurality of feasible solutions that a plurality of initial cargo set pairs are answered constitute is as the input of ant group algorithm, and the step of utilizing the ant group algorithm iterative search to obtain the optimum vanning scheme of the described goods of waiting to case can also may further comprise the steps:

The initial information element of lade is treated in calculating;

Calculate ant quantity, the controlled variable of initialization ant group algorithm according to the described kind of lade for the treatment of;

Placing every ant at random treats on the initial position of lade in each feasible solution;

According to described qualitative restrain, the center of gravity constraint for the treatment of lade, judge and describedly treat whether a class goods on the ant place initial position in the lade can pack into the container, and after reading in the described barycentric coordinates for the treatment of lade, the described class goods in can packing into the container is carried out layout optimization by the placement direction constraint;

Search for current class according to state transition probability and treat that next class of lade treats lade, and control described ant according to Search Results and place described next class to treat on the lade, according to described qualitative restrain, the center of gravity constraint for the treatment of lade, judge and describedly treat whether a class goods on the ant position in the lade can pack into the container, and the described class goods in can packing into the container is carried out layout optimization by the placement direction constraint;

Record Search Results after a cyclic search of the set of feasible solution is finished, and according to the described pheromones of pheromones renewal model modification, carry out the cyclic search next time of the set of feasible solution, when cycle index equates with the number of described a plurality of feasible solutions, finish ant group algorithm, output obtains the optimum vanning scheme of the described goods of waiting to case.

Further, described calculating treats that the step of the initial information element of lade can be expressed as:

τ _ij(0)＝τ _C+τ _G

Wherein, τ _Ij(0) is the described initial information element for the treatment of lade, τ _CBe a default pheromones constant, τ _GSatisfy τ _G=(∑ l _iW _iH _i/ L _jW _jH _j) * 100%, l _iBe the length of i class goods, w _iBe the width of i class goods, h _iIt is the height of i class goods; L _jBe the length of container, W _jBe the width of container, H _jHeight for container.

Further, described judge describedly treat the step of the class goods on the ant place initial position in the lade in whether can packing into the container, and/or judge that the described step of the class goods on the ant position in the lade in whether can packing into the container for the treatment of can be expressed as:

C _q＝∑g≤G

$C_g=\mathrm{cx}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·x_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cx}2,\mathrm{cy}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·y_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cy}2,\mathrm{cz}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·z_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cz}2$

Wherein, ∑ g is the weight sum of the class goods on the ant place initial position for the treatment of in the lade, and/or treats the weight sum of the class goods on the ant position in the lade; G is the weight sum of the container goods that can load; [cx1, cx2], [cy1, cy2], [cz1, cz2] are respectively the boundary value of container in the axial center of gravity safe range of x, y, z; m _iFor treating the quality of the i class goods on the ant place initial position in the lade; (xi, yi zi) are the barycentric coordinates for the treatment of lade.

Further, described state transition probability can be expressed as:

${P_{\mathrm{ij}}}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{{\text{\τ}}_j}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}{\underset{s\∈\mathrm{allowed}}{\mathrm{\Σ}}{{\mathrm{\τ}}_s}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}j\∈& \mathrm{allowe}{d}_k\\ 0& \mathrm{otherwise}\end{array}\right.$

Wherein,

Be heuristic function, dz (j) is for treating the bearing capacity of lade j; v _jIt is the volume for the treatment of lade j; η _Ij ^β(t) search the inspiration degree for the treatment of lade j for ant from treating lade i; Allowed _k=(1,2 ... n)-tabu _kRepresent that ant k is allowed to the lade of placing for the treatment of, tabu next time _kConstantly searched at t and before this loop ends, forbidden the taboo table for the treatment of lade that visits again for having write down ant k; τ _j ^α(t) be pheromones intensity on the goods j;

Described pheromones is upgraded model representation:

τ _j(t+1)＝ρ·τ _j(t)+Δτ _j(t，t+1)

$\mathrm{\Δ}{\mathrm{\τ}}_j(t,t+1)=\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\mathrm{\Δ}{{\mathrm{\τ}}_j}^k(t,t+1)$

Wherein,

Being ant k constantly, (t t+1) stays pheromones amount on the goods j; (1-ρ) is the volatility coefficient of pheromones; f ^k(t) be the container loading rate that ant k searches constantly at t.

The present invention also provides a kind of system of three-dimensional container loading layout optimization, and described system comprises:

Filial generation goods collection generation module is used to treat the vanning goods and encodes, and generates and waits a plurality of initial cargo collection of goods of casing, and utilize genetic algorithm to generate the filial generation goods collection of described initial cargo collection;

The fitness value calculation module is used to calculate each individual fitness value of filial generation cargo consolidation and each individual fitness value of parent cargo consolidation thereof that described filial generation goods collection generation module generates;

Judge module is used to judge whether relative each individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation that described fitness value calculation module calculates increases;

The feasible solution output module, be used for when described judge module is judged each individual fitness value sum of filial generation cargo consolidation that described fitness value calculation module calculates each individual fitness value sum of its parent cargo consolidation is no longer increased relatively the output feasible solution;

The optimal case search module is used for set that a plurality of feasible solutions with the output of described feasible solution output module the constitute input as ant group algorithm, utilizes ant group algorithm iterative search obtain waiting the casing optimum vanning scheme of goods.

The method of three-dimensional container loading layout optimization provided by the invention is utilized the quick at random search capability of genetic algorithm, potential concurrency, global convergence is sought one group of rough feasible solution in solution space, organize of the input of rough feasible solution with this afterwards as ant group algorithm, utilize the positive feedback mechanism of ant group algorithm, the ability of concurrency and search better solutions is tried to achieve the optimal case of vanning, thereby realized that genetic algorithm and ant group algorithm are in the fusion that solves on the container loading location problem, avoided the single algorithm of existing employing to solve the defective of three-dimensional container loading location problem, when taking into account ability of searching optimum, taken into account the several important restrictions conditions that influence efficiency of loading, applicability is good.

Description of drawings

Fig. 1 is the process flow diagram of the method for three-dimensional container loading layout optimization provided by the invention;

Fig. 2 is a kind of flowchart of the method for three-dimensional container loading layout optimization provided by the invention;

Fig. 3 is the schematic diagram of the system of three-dimensional container loading layout optimization provided by the invention.

Embodiment

In order to make purpose of the present invention, technical scheme and advantage clearer,, the present invention is further elaborated below in conjunction with drawings and Examples.Should be appreciated that specific embodiment described herein only in order to explanation the present invention, and be not used in qualification the present invention.

Fig. 1 shows the flow process of the method for three-dimensional container loading layout optimization provided by the invention.

In step S101, treat the vanning goods and encode, generate and wait a plurality of initial cargo collection of goods of casing, and utilize genetic algorithm to generate the filial generation goods collection of this initial cargo collection.

Particularly, treating the vanning goods encodes, generation wait the to case step of a plurality of initial cargo collection of goods can may further comprise the steps: read in goods information to be cased, treat the vanning goods by placement direction constraint and integer coding scheme and encode, the initial population that generates at random by the size of goods to be installed is as the initial cargo collection.

Traditional genetic algorithm is not suitable in the container loading problem the binary coding form that the encoding scheme of separating adopts, for this reason, among the present invention, number as encoding gene with kind, so that similar goods is placed in together when vanning as far as possible, avoid producing too much idle space, help improving the space availability ratio of container loading.

More specifically, the present invention treats the vanning goods by sequence of natural numbers and is numbered, and the same kind cargo number is identical, dissimilar cargo number differences.Each initial population S that generates at random can be expressed as the symbol string that a code length is 2n, and satisfies S={s ₁, s ₂..., s _i..., s _n, s _N+1..., s _N+i, ..., s _2n, wherein, n represents the species number of goods to be installed, gene s ₁To s _nBe integer, represent the kind numbering of goods respectively, gene s _N+1To s _2nRepresent the placement direction constraint of goods to be installed.For example, suppose that the placement direction constraint of goods has four kinds, and satisfy:

$\left\{\begin{array}{c}{l}_i//L_j,w_i//W_j,h_i//H_j\mathrm{rit}=1\\ w_i//L_j,l_i//W_j,h_i//H_j\mathrm{rit}=2\\ l_i//W_j,h_i//L_j,w_i//H_j\mathrm{rit}=3\\ h_i//W_j,l_i//L_j,w_i//H_j\mathrm{rit}=4\end{array}\right.,$

Wherein, l _i, w _i, h _iBe respectively the length, width of i class goods, highly; L _j, W _j, H _jBe respectively the length, width of container, highly; Rit is the placement direction constraint of goods.Hypothesis goods to be cased has three kinds simultaneously, and first kind of goods number is that 5, the second kinds of goods numbers are 6, and the third goods number is 4, the initial population S of Sheng Chenging at random, promptly initial cargo collection P (0) satisfies: S={1,1,1,1,1,2,2,2,2,2,2,3,3,3,3,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2}, that is: goods is by first kind earlier, second kind again, the third order is cased at last, and the placement direction constraint when the placement direction constraint during first kind of goods vanning is encoded to 1, the second kind of goods vanning is encoded to 2, and the placement direction constraint during the third goods vanning is encoded to 2.

Particularly, the step of utilizing genetic algorithm to generate the filial generation goods collection of this initial cargo collection can may further comprise the steps: initialization genetic algorithm controlled variable; Calculate each individual fitness value among the initial cargo collection P (0) according to fitness function; According to fitness value that calculates and the variation probability that prestores, be the basis iterative processing of selecting, intersect, make a variation with initial cargo collection P (0), obtain a plurality of continuous filial generation goods collection P (t) of initial cargo collection P (0), wherein, t is a filial generation algebraically.

More specifically, the genetic algorithm controlled variable comprise be not limited to algebraically counter t, select probability P e, variation probability P m; Fitness function F satisfies F=(l _iW _iH _i/ L _jW _jH _j) * 100%, can it influence the speed of convergence of genetic algorithm and find optimum solution, big more according to its fitness value that obtains, and it is selected that to enter the probability in next son generation big more.

Among the present invention, has higher global convergence and current optimum individual is not intersected in order to ensure genetic algorithm, genetic manipulations such as variation destroy, select to handle optimum conversation strategy of employing and roulette back-and-forth method, at this moment, according to fitness value that calculates and the variation probability that prestores, with initial cargo collection P (0) is that the basis is selected, intersect, variation is handled, the step that iteration obtains a plurality of continuous filial generation goods collection P (t) of initial cargo collection P (0) can may further comprise the steps: according to the ideal adaptation degree value that calculates and optimum conversation strategy and roulette back-and-forth method, determine the selection probability of each individuality among the initial cargo collection P (0); Selection probability according to each individuality of determining is selected two father's individualities in initial cargo collection P (0); According to the variation probability that prestores to two father's individualities selecting make a variation handle or cross processing after, two father's individualities after handling are inserted into next continuous filial generation cargo consolidation of initial cargo collection P (0), and calculate next each individual fitness value of continuous filial generation cargo consolidation of initial cargo collection P (0), repeat above-mentioned steps and iteration obtains a plurality of continuous filial generation goods collection P (t) of initial cargo collection P (0).

For instance, suppose that type of merchandize to be cased is M, the selection probability is Pi, introduces a stochastic variable r, then selects two father's individualities according to the selection probability of each individuality of determining in initial cargo collection P (0); According to the variation probability that prestores to two father's individualities selecting make a variation handle or cross processing after, the step that two father's individualities after handling is inserted into next continuous filial generation cargo consolidation of initial cargo collection P (0) can be expressed as:

In step S102, calculate filial generation each individual fitness value of cargo consolidation and each individual fitness value of parent cargo consolidation thereof, when relative each the individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation no longer increases, corresponding filial generation goods collection is carried out the individuality decoding, obtain feasible solution; And when relative each the individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation continues to increase, return according to fitness value that calculates and the variation probability that prestores, with initial cargo collection P (0) is the basis processing of selecting, intersect, make a variation, and iteration obtains the step of a plurality of continuous filial generation goods collection P (t) of initial cargo collection P (0).

Owing to the be connected algorithm efficiency that opportunity directly influence merge after of genetic algorithm with ant group algorithm, genetic algorithm finishes then to be difficult to too early search and obtains optimal case, genetic algorithm finished party and caused merging the prolongation of back algorithm search execution time, for this reason, the present invention has adopted and has judged that continuously a filial generation goods collection divides other individual fitness value sum dynamic approach whether relative its parent goods collection increases to realize that the best of genetic algorithm and ant group algorithm merges.If each individual fitness value sum of filial generation cargo consolidation no longer increases, the optimization speed that genetic algorithm then is described has reached the highest, adopt genetic algorithm to carry out interative computation if continue this moment, can cause the meaningless redundant iteration of algorithm, reduce algorithm efficiency, therefore, the present invention is each individual fitness value sum of filial generation cargo consolidation relative each individual fitness value sum of its parent cargo consolidation opportunity of no longer increasing, as the opportunity that is connected of genetic algorithm with ant family algorithm.

Wherein, corresponding filial generation goods collection is carried out individuality decoding, the step that obtains feasible solution specifically may further comprise the steps: take out a goods to be cased in turn from corresponding filial generation cargo consolidation, calculate the case volume of goods of waiting of taking out according to waiting of the taking out goods information of casing; According to vanning constraint condition V '+l _iW _iH _i≤ V judge to take out waits to case during whether goods can pack into the container, wherein, and the measurement of cargo of V ' for having packed into the container, l _iW _iH _iBe the case volume of goods of waiting of taking out, V is the useful volume of container; According to judged result, when the goods to be cased that judge to take out can pack into the container when middle, the case volume of goods of waiting of taking out is added among the measurement of cargo V ' that has packed into the container, when the goods to be cased that judge to take out cannot pack into the container when middle, take out next goods to be cased in turn from corresponding filial generation cargo consolidation, case volume and the measurement of cargo sum that has packed into the container of goods during greater than the useful volume of container when corresponding filial generation cargo consolidation goods all to be cased waiting of having got or taken out, obtain the feasible solution after corresponding filial generation goods collection transforms.

For instance, satisfy as corresponding filial generation goods collection P (t): P (t)={ 1,1,1,1,1,2,2,2,2,2,2,3,3,3,3,1,1,1,1,1,2,2,2,2,2,2,2,2,2, during 2}, if according to judged result, first kind of goods packed four into, second kind of goods packed six into, and the third goods is packed three into, and the feasible solution T that then obtains after corresponding filial generation goods collection P (t) transforms satisfies T={1,1,1,1,0,2,2,2,2,2,2,3,3,3,0,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2}.

In step S103, the set that a plurality of feasible solutions that a plurality of initial cargo set pairs are answered constitute is as the input of ant group algorithm, utilizes ant group algorithm iterative search obtain waiting the casing optimum vanning scheme of goods.

Particularly, step S103 can may further comprise the steps: calculate the initial information element for the treatment of lade; Calculate ant quantity, the controlled variable of initialization ant group algorithm according to the kind for the treatment of lade; Place every ant at random and treat in each feasible solution on the initial position of lade, initial position can be the identical goods of coding in each feasible solution; According to the qualitative restrain Cq that treats lade, center of gravity constraint Cg, in judging whether a class goods on the ant place initial position treat in the lade can pack into the container, and after reading in the barycentric coordinates for the treatment of lade, rit carries out layout optimization to such goods in can packing into the container by the placement direction constraint; Search for current class according to state transition probability and treat that next class of lade treats lade, and place next class to treat on the lade according to Search Results control ant, according to the qualitative restrain Cq that treats lade, center of gravity constraint Cg, in judging whether a class goods on the ant position treat in the lade can pack into the container, and such goods in can packing into the container is carried out layout optimization by placement direction constraint rit; Record Search Results after a cyclic search of the set of feasible solution is finished, and according to pheromones renewal model modification pheromones, carry out the cyclic search next time of the set of feasible solution, when cycle index equates with the number of a plurality of feasible solutions, finish ant group algorithm, output obtain waiting the casing optimum vanning scheme of goods.

Wherein, the controlled variable of ant group algorithm includes but not limited to the controlled variable α of ant group algorithm, β, the pheromones increment

The pheromones constant tau _C, pheromones volatilization factor ρ, taboo table.

More specifically, calculating treats that the step of the initial information element of lade can be expressed as: τ _Ij(0)=τ _C+ τ _G, wherein, τ _Ij(0) for treating the initial information element of lade, τ _CBe a default pheromones constant, τ _GBe the pheromones value of calculating according to the genetic algorithm Search Results, and satisfy τ _G=(∑ l _iW _iH _i/ L _jW _jH _j) * 100%.

More specifically, judge a class goods on the ant place initial position treat in the lade step in whether can packing into the container, and/or judge that a class goods on the ant position for the treatment of in the lade step in whether can packing into the container can be expressed as:

C _q＝∑g≤G

$C_g=\mathrm{cx}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·x_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cx}2,\mathrm{cy}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·y_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cy}2,\mathrm{cz}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·z_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cz}2$

Wherein, ∑ g is the weight sum of the class goods on the ant place initial position for the treatment of in the lade, and/or treats the weight sum of the class goods on the ant position in the lade; G is the weight sum of the container goods that can load; [cx1, cx2], [cy1, cy2], [cz1, cz2] are respectively the boundary value of container in the axial center of gravity safe range of x, y, z; m _iFor treating the quality of the i class goods on the ant place initial position in the lade; (xi, yi zi) are the barycentric coordinates for the treatment of lade.

More specifically, state transition probability can be expressed as:

${P_{\mathrm{ij}}}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{{\text{\τ}}_j}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}{\underset{s\∈\mathrm{allowed}}{\mathrm{\Σ}}{{\mathrm{\τ}}_s}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}j\∈& \mathrm{allowe}{d}_k\\ 0& \mathrm{otherwise}\end{array}\right.$

Wherein,

Be heuristic function, dz (j) is for treating the bearing capacity of lade j (j is a positive integer); v _jIt is the volume for the treatment of lade j; η _Ij ^β(t) search the inspiration degree for the treatment of lade j for ant from treating lade i, η _Ij ^β(t) the rule that is provided with is directly proportional and is inversely proportional to the bearing capacity for the treatment of lade with the volume for the treatment of lade j, places bottom and is damaged by pressure than low-density cargo avoiding; Allowed _k=(1,2 ... n)-tabu _kRepresent that ant k (k is a positive integer) is allowed to the lade of placing for the treatment of, tabu next time _kConstantly searched at t and before this loop ends, forbidden the taboo table for the treatment of lade that visits again for having write down ant k; τ _j ^α(t) be pheromones intensity on the goods j, intensity is big more, and the probability of selected loading is big more.

More specifically, pheromones more new model can be expressed as:

τ _j(t+1)＝ρ·τ _j(t)+Δτ _j(t，t+1)

$\mathrm{\Δ}{\mathrm{\τ}}_j(t,t+1)=\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\mathrm{\Δ}{{\mathrm{\τ}}_j}^k(t,t+1)$

Wherein,

Being ant k constantly, (t t+1) stays pheromones amount on the goods j; (1-ρ) is the volatility coefficient of pheromones; f ^k(t) be the container loading rate that ant k searches constantly at t.

Fig. 2 shows a kind of execution flow process of the method for three-dimensional container loading layout optimization provided by the invention.

In step S201, treat vanning goods coding, generate filial generation goods collection;

In step S202, calculate cargo consolidation ideal adaptation degree value;

In step S203, judge whether each individual fitness value sum of cargo consolidation increases, be execution in step S206 then, otherwise execution in step S204;

In step S204, to the processing of selecting, intersect, make a variation of goods collection;

In step S205, generate next son for the goods collection, and return step S202;

In step S206, individual decoding obtains the set of feasible solution;

In step S207, ant is placed inhomogeneity treat lade, and treat the device goods according to constraint condition and carry out layout optimization;

In step S208, place next class to treat lade ant according to state transition probability;

In step S209, the lastest imformation element;

In step S210, judge whether cycle index equals the feasible solution number, be execution in step S211 then, otherwise execution in step S212;

In step S211, export optimum vanning scheme;

In step S212, recursive iteration, and return step S207.

Fig. 3 shows the principle of the system of three-dimensional container loading layout optimization provided by the invention.

The system of three-dimensional container loading layout optimization provided by the invention comprises: filial generation goods collection generation module 11, being used to treat the vanning goods encodes, the a plurality of initial cargo collection of goods of casing are waited in generation, and utilize genetic algorithm to generate the filial generation goods collection of this initial cargo collection; Fitness value calculation module 12 is used to calculate each individual fitness value of filial generation cargo consolidation and each individual fitness value of parent cargo consolidation thereof that filial generation goods collection generation module 11 generates; Judge module 13 is used to judge whether relative each individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation that fitness value calculation module 12 calculates increases; Feasible solution output module 14 is used for when judge module 13 is judged each individual fitness value sum of filial generation cargo consolidation that fitness value calculation modules 12 calculate each individual fitness value sum of its parent cargo consolidation is no longer increased relatively the output feasible solution; Optimal case search module 15 is used for set that a plurality of feasible solutions with feasible solution output module 14 output the constitute input as ant group algorithm, utilizes ant group algorithm iterative search obtain waiting the casing optimum vanning scheme of goods.

Wherein, as above the method part is described for filial generation goods collection generation module 11, fitness value calculation module 12, judge module 13 and 15 fens other execution in step of optimal case search module, does not repeat them here.

The method of three-dimensional container loading layout optimization provided by the invention is utilized the quick at random search capability of genetic algorithm, potential concurrency, global convergence is sought one group of rough feasible solution in solution space, organize of the input of rough feasible solution with this afterwards as ant group algorithm, utilize the positive feedback mechanism of ant group algorithm, the ability of concurrency and search better solutions is tried to achieve the optimal case of vanning, thereby realized that genetic algorithm and ant group algorithm are in the fusion that solves on the container loading location problem, avoided the single algorithm of existing employing to solve the defective of three-dimensional container loading location problem, when taking into account ability of searching optimum, taken into account the several important restrictions conditions that influence efficiency of loading, applicability is good.

The above; only be the preferable embodiment of the present invention; but protection scope of the present invention is not limited thereto; anyly be familiar with those skilled in the art in the technical scope that the present invention discloses; be equal to replacement or change according to technical scheme of the present invention and inventive concept thereof, all should be encompassed within protection scope of the present invention.

Claims (10)

1. the method for a three-dimensional container loading layout optimization is characterized in that, said method comprising the steps of:

Treat the vanning goods and encode, generate and wait a plurality of initial cargo collection of goods of casing, and utilize genetic algorithm to generate the filial generation goods collection of described initial cargo collection;

Calculate filial generation each individual fitness value of cargo consolidation and each individual fitness value of parent cargo consolidation thereof, when relative each the individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation no longer increases, corresponding filial generation goods collection is carried out the individuality decoding, obtain feasible solution;

The set that a plurality of feasible solutions that a plurality of initial cargo set pairs are answered constitute is as the input of ant group algorithm, utilizes the ant group algorithm iterative search to obtain the optimum vanning scheme of the described goods of waiting to case.

2. the method for three-dimensional container loading layout optimization as claimed in claim 1 is characterized in that, the described step of the filial generation goods collection that genetic algorithm generates described initial cargo collection of utilizing is further comprising the steps of:

Initialization genetic algorithm controlled variable;

Calculate described initial cargo according to fitness function and concentrate each individual fitness value;

According to described fitness value that calculates and the variation probability that prestores, be the basis iterative processing of selecting, intersect, make a variation with described initial cargo collection, obtain a plurality of continuous filial generation goods collection of described initial cargo collection.

3. the method for three-dimensional container loading layout optimization as claimed in claim 2 is characterized in that, described fitness function satisfies relational expression:

F＝(l _i·w _i·h _i/L _j·W _j·H _j)×100％

Wherein, F is a fitness function; l _iBe the length of i class goods, w _iBe the width of i class goods, h _iIt is the height of i class goods; L _jBe the length of container, W _jBe the width of container, H _jHeight for container.

4. the method for three-dimensional container loading layout optimization as claimed in claim 2, it is characterized in that, described selection is handled and is adopted optimum conversation strategy and roulette back-and-forth method, described fitness value that described basis calculates and the variation probability that prestores, with described initial cargo collection is the basis iterative processing of selecting, intersect, make a variation, and the step of a plurality of continuous filial generation goods collection that obtains described initial cargo collection is further comprising the steps of:

According to the ideal adaptation degree value that calculates and optimum conversation strategy and roulette back-and-forth method, determine the selection probability of concentrated each individuality of described initial cargo;

Selection probability according to described each individuality of determining is selected two father's individualities in described initial cargo collection;

According to the variation probability that prestores described two father's individualities of selecting are made a variation and to handle or cross processing, two father's individualities after handling are inserted into next continuous filial generation cargo consolidation of described initial cargo collection, and calculate next each individual fitness value of continuous filial generation cargo consolidation of described initial cargo collection;

According to each the individual fitness value of current filial generation cargo consolidation and optimum conversation strategy and the roulette back-and-forth method that calculate, determine the selection probability of current each individuality of filial generation cargo consolidation;

Selection probability according to current each individuality of filial generation cargo consolidation of determining is selected two father's individualities in current filial generation goods collection;

According to the variation probability that prestores described two father's individualities of selecting are made a variation and to handle or cross processing, two father's individualities after handling are inserted into next continuous filial generation cargo consolidation of current filial generation goods collection, thereby iteration obtains a plurality of continuous filial generation goods collection of described initial cargo collection.

5. the method for three-dimensional container loading layout optimization as claimed in claim 1 is characterized in that, described corresponding filial generation goods collection is carried out individuality decoding, and the step that obtains feasible solution is further comprising the steps of:

Take out a goods to be cased in turn from corresponding filial generation cargo consolidation, calculate take out described wait the to case volume of goods according to the described goods information of waiting to case that takes out;

According to vanning constraint condition V '+l _iW _iH _iDuring≤V judges whether the described goods of waiting to case that takes out can pack into the container, wherein, the measurement of cargo of V ' for having packed into the container, l _iW _iH _iBe the volume of the described goods of waiting to case that takes out, V is the useful volume of container;

According to judged result, when the goods described to be cased that judge to take out can pack into the container when middle, the volume of the described goods of waiting to case that takes out is added among the described measurement of cargo V ' that has packed into the container, when the goods described to be cased that judge to take out cannot pack into the container when middle, take out next goods to be cased in turn from corresponding filial generation cargo consolidation, case volume and the measurement of cargo sum that has packed into the container of goods during greater than the useful volume of described container when corresponding filial generation cargo consolidation goods all to be cased waiting of having got or taken out, obtain the feasible solution after corresponding filial generation goods collection transforms.

6. the method for three-dimensional container loading layout optimization as claimed in claim 1, it is characterized in that, the set that the described a plurality of feasible solutions that a plurality of initial cargo set pairs are answered constitute is as the input of ant group algorithm, and it is further comprising the steps of to utilize the ant group algorithm iterative search to obtain the step of optimum vanning scheme of the described goods of waiting to case:

The initial information element of lade is treated in calculating;

Calculate ant quantity, the controlled variable of initialization ant group algorithm according to the described kind of lade for the treatment of;

Placing every ant at random treats on the initial position of lade in each feasible solution;

According to described qualitative restrain, the center of gravity constraint for the treatment of lade, judge and describedly treat whether a class goods on the ant place initial position in the lade can pack into the container, and after reading in the described barycentric coordinates for the treatment of lade, the described class goods in can packing into the container is carried out layout optimization by the placement direction constraint;

Search for current class according to state transition probability and treat that next class of lade treats lade, and control described ant according to Search Results and place described next class to treat on the lade, according to described qualitative restrain, the center of gravity constraint for the treatment of lade, judge and describedly treat whether a class goods on the ant position in the lade can pack into the container, and the described class goods in can packing into the container is carried out layout optimization by the placement direction constraint;

Record Search Results after a cyclic search of the set of feasible solution is finished, and according to the described pheromones of pheromones renewal model modification, carry out the cyclic search next time of feasible solution set, when cycle index equates with the number of described a plurality of feasible solutions, finish ant group algorithm, output obtains the optimum vanning scheme of the described goods of waiting to case.

7. the method for three-dimensional container loading layout optimization as claimed in claim 6 is characterized in that, described calculating treats that the step of the initial information element of lade is expressed as:

τ _ij(0)＝τ _C+τ _G

Wherein, τ _Ij(0) is the described initial information element for the treatment of lade, τ _CBe a default pheromones constant, τ _GSatisfy τ _G=(∑ l _iW _iH _i/ L _jW _jH _j) * 100%, l _iBe the length of i class goods, w _iBe the width of i class goods, h _iIt is the height of i class goods; L _jBe the length of container, W _jBe the width of container, H _jHeight for container.

8. the method for three-dimensional container loading layout optimization as claimed in claim 6, it is characterized in that, described judge describedly treat the step of the class goods on the ant place initial position in the lade in whether can packing into the container, and/or judge that the described step of the class goods on the ant position in the lade in whether can packing into the container for the treatment of is expressed as:

C _q＝∑g≤G

$C_g=\mathrm{cx}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·x_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cx}2,\mathrm{cy}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·y_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cy}2,\mathrm{cz}1\≤\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i\·z_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i}\≤\mathrm{cz}2$

Wherein, ∑ g is the weight sum of the class goods on the ant place initial position for the treatment of in the lade, and/or treats the weight sum of the class goods on the ant position in the lade; G is the weight sum of the container goods that can load; [cx1, cx2], [cy1, cy2], [cz1, cz2] are respectively the boundary value of container in the axial center of gravity safe range of x, y, z; m _iFor treating the quality of the i class goods on the ant place initial position in the lade; (xi, yi zi) are the barycentric coordinates for the treatment of lade.

9. the method for three-dimensional container loading layout optimization as claimed in claim 6 is characterized in that, described state transition probability is expressed as:

${P_{\mathrm{ij}}}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{{\text{\τ}}_j}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}{\underset{s\∈\mathrm{allowed}}{\mathrm{\Σ}}{{\mathrm{\τ}}_s}^{\mathrm{\α}}\left(t\right)\·{{\mathrm{\η}}_{\mathrm{ij}}}^{\mathrm{\β}}\left(t\right)}j\∈& \mathrm{allowe}{d}_k\\ 0& \mathrm{otherwise}\end{array}\right.$

Wherein,

Be heuristic function, dz (j) is for treating the bearing capacity of lade j; v _jIt is the volume for the treatment of lade j; η _Ij ^β(t) search the inspiration degree for the treatment of lade j for ant from treating lade i; Allowed _k=(1,2 ... n)-tabu _kRepresent that ant k is allowed to the lade of placing for the treatment of, tabu next time _kConstantly searched at t and before this loop ends, forbidden the taboo table for the treatment of lade that visits again for having write down ant k; τ _j ^α(t) be pheromones intensity on the goods j;

Described pheromones is upgraded model representation:

τ _j(t+1)＝ρ·τ _j(t)+Δτ _j(t，t+1)

$\mathrm{\Δ}{\mathrm{\τ}}_j(t,t+1)=\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\mathrm{\Δ}{{\mathrm{\τ}}_j}^k(t,t+1)$

Wherein,

Being ant k constantly, (t t+1) stays pheromones amount on the goods j; (1-ρ) is the volatility coefficient of pheromones; f ^k(t) be the container loading rate that ant k searches constantly at t.

10. the system of a three-dimensional container loading layout optimization is characterized in that, described system comprises:

Filial generation goods collection generation module is used to treat the vanning goods and encodes, and generates and waits a plurality of initial cargo collection of goods of casing, and utilize genetic algorithm to generate the filial generation goods collection of described initial cargo collection;

The fitness value calculation module is used to calculate each individual fitness value of filial generation cargo consolidation and each individual fitness value of parent cargo consolidation thereof that described filial generation goods collection generation module generates;

Judge module is used to judge whether relative each individual fitness value sum of its parent cargo consolidation of each individual fitness value sum of filial generation cargo consolidation that described fitness value calculation module calculates increases;

The feasible solution output module, be used for when described judge module is judged each individual fitness value sum of filial generation cargo consolidation that described fitness value calculation module calculates each individual fitness value sum of its parent cargo consolidation is no longer increased relatively the output feasible solution;

The optimal case search module is used for set that a plurality of feasible solutions with the output of described feasible solution output module the constitute input as ant group algorithm, utilizes ant group algorithm iterative search obtain waiting the casing optimum vanning scheme of goods.

Images