CN110378644A - A kind of harbour dangerous material container space distribution model and algorithm - Google Patents
A kind of harbour dangerous material container space distribution model and algorithm Download PDFInfo
- Publication number
- CN110378644A CN110378644A CN201910234955.6A CN201910234955A CN110378644A CN 110378644 A CN110378644 A CN 110378644A CN 201910234955 A CN201910234955 A CN 201910234955A CN 110378644 A CN110378644 A CN 110378644A
- Authority
- CN
- China
- Prior art keywords
- individual
- dangerous material
- container
- stockpiling
- population
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06313—Resource planning in a project environment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Development Economics (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Biodiversity & Conservation Biology (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of harbour dangerous material container space distribution model and algorithms, belong to storage space allocation model and algorithm field, the following steps are included: S1: integrating the Codes and Standards during dangerous material stockpiling, S2: the specific requirement of reservoir area and number of plies limitation can be stored up by investigating all kinds of dangerous material, S3: building harbour dangerous material container stores up Model for Multi-Objective Optimization, S4: setting multiple target improved adaptive GA-IAGA, S5: harbour dangerous material container stockpiling Model for Multi-Objective Optimization is solved, obtains optimal dangerous material container space stacked arrangement.The present invention is under conditions of guaranteeing dangerous cargo warehouse general safety, optimization stockpiling case position in the shortest possible time provides a new point of penetration for harbour dangerous material container management, and reduces the planning container space time, promote the competition soft power and Hard Power at economic benefits and harbour.
Description
Technical field
The present invention relates to storage space allocation model and algorithm, in particular to the foundation of dangerous material container stockpiling Optimized model with
And a kind of multi-objective problem derivation algorithm of improved dangerous material container stockpiling Optimized model.
Background technique
With economic stable development, logistics is quickly grown as third party's profit source, therefore enterprise is looked forward to be promoted
Industry competitive strength, more visual angles start to turn to the potentiality for excavating logistics field.Modern international logistics is mostly to rely on to have transported by sea
At, the container as one of international freight transport mode starts to develop to enlargement, high speed direction, this undoubtedly gives harbour
Huge pressure is brought, therefore the safety problem at harbour also becomes increasingly conspicuous, especially the safety of dangerous material.
The work such as the handling for carrying out dangerous material container, transhipment, unpacking, repairing can be concentrated on stockyard, for convenience
Determine container position information, stockyard usually marks " case position " line, puts a number on, number is by case area, Bei Wei, column and floor come table
Showing, " case area-shellfish position-row-floor " unique identification can be used in each container stockpiling case position, and each stockyard is divided into several areas,
If each dividing into dried scallop, each shellfish is divided into several columns, and each column is made of several layers container, and each container is generally standard
20 ruler cabinet containers.And shellfish position in vertical direction stack container be known as stack, the container in stack be usually it is advanced
Bottom, laggard store in top, possible stack is a container, it is also possible to have multiple containers, wherein columns and layer
Number is all continuous number, and shellfish number is usually to be made of odd number, if indicate shellfish number is an even number, illustrates that container is 40
Ruler cabinet container needs to account for two positions.
Harbour dangerous cargo warehouse business is divided into two kinds, and one is inlet box business, and one is EXPORT CARTON business, two kinds of business are equal
It is using dangerous material Container Yard as platform, but process is different, and refering to attached drawing 1-2, two ways stockpiling optimization is all base
In the container of regular general cargo, rather than the dangerous material container with nature of danger, therefore lack and meet dangerous material property
Restrictive condition.
Existing harbour dangerous material container space distribution model and algorithm, main enlightenment formula method and meta-heuristic side
Method both, due to needing the objective functions that optimize, and these target letters there are multiple when dangerous material container, which is stored up, to be optimized
It is not fully function in the same direction between number, frequently can lead to the drop of other objective function performances to the optimization of wherein some function
It is low, thus make simultaneously all targets be optimal state be it is impossible, dangerous material container management at this stage is often
Referring to container management, the optimum management of a certain specific objective is only valued, such as similar ship container stacking is to together, and ignore
The Synchronous fluorimetries of other targets.
Summary of the invention
It is above-mentioned to solve the purpose of the present invention is to provide a kind of harbour dangerous material container space distribution model and algorithm
The problem of being proposed in background technique.
To achieve the above object, the invention provides the following technical scheme: a kind of harbour dangerous material container space distributes mould
Type and algorithm, comprising the following steps:
S1: carrying out analytical integration to dangerous material and summarize, obtain Codes and Standards of dangerous material container during stockpiling,
Such as classify stockpiling, stockpiling height limitation, as one of the restrictive condition of multiple-objection optimization later；
S2: investigating the stockpiling situation of harbour dangerous cargo warehouse, and reservoir area and number of plies limit can be stored up by obtaining all kinds of dangerous material
The specific requirement of system；
S3: building harbour dangerous material container stores up Model for Multi-Objective Optimization；
S4: setting multiple target improved adaptive GA-IAGA；
S5: harbour dangerous material container stockpiling Model for Multi-Objective Optimization is solved, obtains optimal dangerous material packaging
Case case position stacked arrangement.
Preferably, the method that analytical integration summarizes is carried out to dangerous material in the S1, mainly by referring to such as GB
The relevant laws and regulations of 6944-2012, GB 12268-2012, JT 397-2007 and GB 18218-2009.
Preferably, the building of model considers following hypothesis in the S3: not considering the difference of container size, is 20 rulers
Cabinet；Just for hazardous chemical；March into the arena plan and stockyard stored up container it is known that include dangerous material container type, quality,
It marches into the arena the time, if EXPORT CARTON, then corresponds to known to time for competiton, course line and port of destination.
Preferably, model builds following set in the S3:
All container set, I={ 1,2,3 ..., Ni, and i, j ∈ I；
Container of showing up is total, Ni；
The set of field, K={ 1,2,3 ..., }；
The set of shellfish, B={ 01,03,05 ..., Nb}；
The set of row, R={ 1,2,3 ..., Nr}；
The set of layer, T={ 1,2,3 ..., Nt}；
Model builds following parameter in the S3:
SiExpression is put into specified place according to classification for premise, is 1 when meeting, and is 0 when not meeting；
Ei,jThe case first marched into the arena for premise is indicated in lower layer, is 1 when meeting, is 0 when not meeting；
Oi,jThe case first to appear on the scene for premise is indicated on upper layer, is 1 when meeting, is 0 when not meeting；
Wi,jIndicate same for premise voyage, weight case upper, be 1 when meeting, be 0 when not meeting；
Ob,r,tIndicate occupied case position；
Cb,r,tIndicate the transportation cost of different shellfishes；
Model builds following decision variable in the S3:
xI, b, r, t, otherwise it is 0 that container i, which is 1 when being placed to t layers of r column of b shellfish,；
Y, i and j are 1 when being assigned to two case positions, are otherwise 0；
zeIf the case that the stockpiling of container i and container j first march into the arena for premise in lower layer, is 1 when meeting, it is not inconsistent
It is 0 when conjunction；
zoIf the case that the stockpiling of container i and container j first appear on the scene for premise on upper layer, is 1 when meeting, it is not inconsistent
It is 0 when conjunction；
zwIf the case of the stockpiling of container i and container j voyage same for premise, weight upper, be 1 when meeting, be not inconsistent
It is 0 when conjunction；
qkIndicate the actual mass (unit is ton) of dangerous cargo k stockpiling, QkIt is critical to indicate that dangerous cargo k can store up
Quality (unit is ton), type included in k ∈ subordinate list 1.
Preferably, the multi-objective Model of four targets is mainly made of objective function and constraint condition, objective function
For Min f={ fo, fw, fh, fc, wherein
(1) it is minimum on upper layer to be unsatisfactory for the case first to appear on the scene
fo=∑I, j, b, r, t1 < t2((y-zo)*(t2-t1)) (3-1)
(2) case of regular EXPORT CARTON, same to voyage, weight is unsatisfactory for upper minimum
fw=∑I, j, b, r, t1 < t2((y-zw)*(t2-t1)) (3-2)
(3) stockpiling height is minimized.
fh=∑I, b, r, t(XI, b, r, t*t2) (3-3)
(4) transportation range is nearest.
fc=∑b(CB, r, t*∑B, r, t(XI, b, r, t)) (3-4)
Constraint condition:
Wherein constraint condition (3-5) and (3-6) indicate to be placed on j as i in the following, and i ratio j is early when marching into the arena, zeTake 1, otherwise for
0；(3-7) and (3-8) indicates decision variable y and xi,b,r,tBetween relationship, only when container i and j are respectively allocated to case position
(b,r,t1) and (b, r, t2) when, the value of decision variable y takes 1, is otherwise 0；(3-9) indicates that chest cannot be placed vacantly；(3-
10) indicate that dangerous material are stored in the place that store up；(3-11) indicates that each on-site will have baffle-box position.
Preferably, the setting of multiple target improved adaptive GA-IAGA includes the following steps: in the S3
S1: initial population setting uses in order to which the order of container and specific location more directly show
The mode of real number encodes, i.e., is encoded with the position of container, and a position of the chromosome of an individual indicates heap
Actual position in putting, encoding scheme is as shown in the picture, indicates that the positions of all possible storages after container is shown up have m, and
The distribution position of preceding n expression container and allocation order；
S2: quick non-dominated ranking, steps are as follows:
S201: after generating initial population, to each individual xiAll set a parameter niWith a set si, niIt is this
The number for the solution that individual is dominated by other individuals, that is, the individual amount more outstanding than the individual；siIt is to be dominated by the individual
In other words other groups of individuals in population are exactly group of individuals more worse than the individual；
S202: individual x is calculatediThe value in each optimization object function, that is, bring formula (3-1), (3-2), (3- into
3), in (3-4), the functional value f of all targets has been obtainedo(xi), fw(xi), fh(xi), fc(xi)；
S203: in addition to xiOutside, other individuals x in population is traversedj, similarly, by each individual xjBring a formula into respectively
(3-1), (3-2), (3-3), in (3-4), the functional value f of the individual is obtainedo(xj), fw(xj), fh(xj), fc(xj) comparison two
Value of the individual about all objective functions: if meeting fo(xi) < fo(xj), fw(xi) < fw(xj), fh(xi) < fh(xj), fc
(xi) < fc(xj) then claim individual xiDominate individual xj, dominate set siIn to add individual xj；If meeting fo(xi) > fo
(xj), fw(xi) > fw(xj), fh(xi) > fh(xj), fc(xi) > fc(xj) then claim individual xiBy individual xjIt dominates, by domination parameter
niIncrease by 1；
S204: individual x in population has been traversedjAfterwards, individual x is observediParameter ni: if it is equal to 0, illustrates individual xiIt should not
It is dominated by other individuals in population, it should non-dominant level F be added1；If it is greater than 0, illustrate individual xiNon- branch cannot be incorporated into
With level；
S205: returning in S202, continues to traverse other individuals, complete until traversing；
S206: non-dominant level F is generated according to S202-S2051, wherein individual xiCorresponding set siIn whole
Individual xj, to the parameter n of individual each in setj, n is executed to itj=nj- 1 operation, then judges njValue: if be equal to 0,
Illustrate individual xjIt is currently no longer dominated by other individuals, level F can be added2；If being not equal to 0, explanation cannot be incorporated into layer
Grade F2；
S207: other individuals are looped through, until F1Traversal is complete；
S208: to non-dominant F2Do with the identical operation of S206, until next level be sky；S3: improved to gather around
Degree calculation method is squeezed, steps are as follows:
S301: the individual x in non-dominant level is traversedi, the functional value of some objective function is calculated, is carried out according to functional value
Descending arrangement is set as maximum value to the crowding of the individual on sequence both sides, it is ensured that it is chosen to always；
S302: since the 2nd individual to second-to-last individual until, calculate individual xj crowding before, first sentence
Breaking, whether it is identical with individual before: if identical, no longer calculating individual xjCrowding, but directly will
Its crowded angle value is set as the crowded angle value of same individual, and by individual xjNon-dominant hierarchical value on add population scale, increase
The big non-dominant level of the individual, reduces selected probability, the label of realization redundancy individual, if it is not, then according to such as
Lower crowding formula is calculated:
S303: if there is multiple optimization object functions, return step S301, until not traversing
Until complete all optimization aims；
S4: improved elite retention strategy is judged redundancy individual using label, and is incorporated into interim level, finally
In newly-generated population scale deficiency, it is taken out corresponding redundancy individual and is incorporated to new population, steps are as follows:
S401: parent and filial generation are merged into operation, population scale is made to be reached for 2N；
S402: quick non-dominated ranking is carried out according to S2 to the population after merging, determines each non-dominant level；
S403: to each non-dominant level, the crowding of individual is calculated according to S3；
S404: creation new population P respectively operates each non-dominant level, first to FiIn individual according to gathering around
It squeezes angle value and carries out descending sort, then traverse each of these individual, judge whether the non-dominant level of current individual is greater than
2N: if so, illustrating that this is a redundancy individual, refusal is incorporated into new population P, but puts it into interim level Ftemp
In；If it is not, then continuing the level after traversal；If whole traversals is complete, the scale of new population is less than N, then to interim
Individual in level is ranked up according to crowded angle value, and it is new to be incorporated to that individual is successively then taken out out of interim level according to sequence
Population, and individual first carries out mutation operation to it before being incorporated to new population, until new population scale is N；
S5: genetic manipulation, steps are as follows:
S501: initial population is selected using wheel disc bet method is improved, steps are as follows:
S50101: descending sort is carried out according to crowded angle value from the individual in the population P for needing to carry out genetic manipulation, i.e., first
Ascending sort is carried out according to non-dominant level belonging to individual, individual identical for non-dominant level is carried out according still further to crowding
Descending sort；
S50102: all crowdings are added summation, and calculate select probability:
Wherein, PiFor the probability that ith member in current iteration is selected, diFor the fitness value of ith member；
S50103: generating random number m, probability value be added, until accumulated probability is equal to or more than m；
S502: crossover process, mainly in such a way that any intersects, i.e., simple to intersect, steps are as follows:
S50201: cross-point locations are determined；
S50202: Duplication part is excluded；
S503: mutation operation judges whether to make a variation according to the random number size that rand (0,1) generates, if can be with
Individual chromosome is then encoded the genic value on some upper position and replaced with the genic value on another position of the chromosome by variation
It changes, after the completion of replacement, to examine the legitimacy of individual, if it is illegal, then need to adjust sequence to meet condition, until scheme can
Row.
Compared with prior art, the beneficial effects of the present invention are: the present invention is provided by analysis dangerous material container stockpiling,
Existing Model for Multi-Objective Optimization is improved, proposes that new constraint optimizes, makes it more close to dangerous material heap
Field stocking requirements, are solved using genetic algorithm, avoid locally optimal solution, improve operation efficiency, guarantee dangerous cargo warehouse general safety
Under conditions of, in the shortest possible time optimization stockpiling case position, for harbour dangerous material container management provide one newly cut
Access point, and the planning container space time is reduced, promote the competition soft power and Hard Power at economic benefits and harbour.
Detailed description of the invention
Fig. 1 is that inlet box of the invention flows out figure；
Fig. 2 is EXPORT CARTON flow chart of the invention；
Fig. 3 is pattern function block diagram of the invention；
Fig. 4 is genetic algorithm encoding conceptual scheme of the invention；
Fig. 5 is the area the embodiment of the present invention KD container distribution map；
Fig. 6 is operational flowchart of the invention.
Specific embodiment
The technical scheme in the embodiments of the invention will be clearly and completely described below, it is clear that described implementation
Example is only a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, this field is common
Technical staff's every other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
Embodiment one:
Fig. 3-5, subordinate list 1-6 are please referred to, the present invention provides a kind of technical solution: a kind of harbour dangerous material container space point
With model and algorithm, comprising the following steps:
S1: carrying out analytical integration to dangerous material and summarize, obtain Codes and Standards of dangerous material container during stockpiling,
Such as classify stockpiling, stockpiling height limitation, as one of the restrictive condition of multiple-objection optimization later；
S2: investigating the stockpiling situation of harbour dangerous cargo warehouse, and reservoir area and number of plies limit can be stored up by obtaining all kinds of dangerous material
The specific requirement of system；
S3: building harbour dangerous material container stores up Model for Multi-Objective Optimization；
S4: setting multiple target improved adaptive GA-IAGA；
S5: harbour dangerous material container stockpiling Model for Multi-Objective Optimization is solved, obtains optimal dangerous material packaging
Case case position stacked arrangement.
Firstly, Codes and Standards of dangerous material container during stockpiling are identical as subordinate list 1- subordinate list 4.
1 Classification of Dangerous of subordinate list
The requirement of 2 segregation of dangerous chemicals of subordinate list
3 major hazard source storage area appreciative standard of subordinate list
4 dangerous material container insulation request of subordinate list
Secondly, the stockpiling situation to dangerous cargo warehouse is investigated, receivable hazardous classification, stockyard point are specifically included
Area, every area can store up the classification of dangerous material and the number of plies limitation of field heap；
Then, instance data is obtained, the data of acquisition include container disengaging field time, weight, shipping company with
And hazardous classification etc., as shown in subordinate list 5.
The actual information that 5 43 containers of subordinate list are shown up
It includes the container of 6.1 classes, 8 classes, 9 classes that the present embodiment, which has selected a certain day, by all containers according to import
After case and EXPORT CARTON classification, classify further according to EXPORT CARTON gross weight, the final area KD for determining chest and being arranged in stockyard, highest heap
Deposit 3 layers.In order to facilitate calculating, data are handled, processing rule is as follows:
(1) first the weight of chest is simplified, is indicated with weight grade；
(2) shipping company is english abbreviation, is not suitable for the typing of data, therefore is reduced to number；
(3) time for competiton determines radix according to ETA estimated time of arrival, different further according to the port of destination of same ship, arrives first port of destination
Evening time for competiton, therefore time zone is separated plus decimal place, such as container A and container B, container B arrives first port, therefore
Time for competiton is more late than container A, therefore the container B time for competiton is set to 2.1, by simplified result as shown in subordinate list 6.
The details that 6 43 containers of subordinate list are shown up
Later, changed to obtain this day vacancy case position according to the case position before the area KD in embodiment, as shown in figure 5, wherein 03
It arranges close to channel.
Finally, carrying out instance model calculating, core is first before performing selection operation to whole individuals in population
Non-dominated ranking is executed to it, so that it is determined that all mutual dominance relations of individual in the population and affiliated
Pareto optimal solution set, as basic genetic algorithm calculation, main flow is as follows for operation later: random to generate rule
Mould is the initial population of N, executes quick non-dominated ranking to whole powders, determines non-dominant level, determines individual according to level
Fitness；The set for traversing the non-dominant level of whole generated by initial population, calculates in the same non-dominant level
Then all crowded angle value execute selection, intersection, mutation operation, generate the progeny population that scale is N；Utilize improved elite
Retention strategy merges filial generation and initial population, generates new population；Selection intersection is carried out to the population and mutation operation generates newly
Progeny population stop algorithm if cycle-index has arrived, if not meeting, return to improved elite and retain step.
S1: initial population and parameter setting:
Initial population generation method: being encoded by the way of real number, i.e., is encoded with the position of container, one by one
One position of the chromosome of body indicates actual position in stacking, then by array upset sequence generate at random it is N number of initial
Population, then its validity is verified: meeting chest cannot vacantly place；
Parameter setting: general including population scale (i.e. initial population number), maximum evolutionary generation, crossover probability and variation
Rate.
S2: quick non-dominated ranking
for xiTo P/* traverse all individual * in population/
/ * initialization individual domination set */
ni=0；The non-dominant parameter * of/* initialization individual/
All objective function * of for obj to 4/* traversal/
Cal (m, xi)；The each target function value * of/* calculating/
for xjAll individual * of to P/* traversal/
for obj to m
Cal (m, xj)；
if fo(xi) < fo(xj)&&fh(xi) < fh(xj)&&fw(xi) < fw(xj)&&fc(xi) < fc(xj)
si=si∪{xj}；/ * individual xiDominate individual xj*/
if f°(xi) > fo(xj)&&fh(xi) > fh(xj)&&fw(xi) > fw(xj)&&fc(xi) > fc(xj)
ni=ni+1；/ * individual xiBy individual xjDomination */
if ni=0
I=1；F1=F1∪{xj}；/ * individual xiNot by any other individual dominate */
if
The next level * of/* creation/
for xi to Flayer/ * traverse all individual * in level/
for xj to si/ * traverses xjAll individual * in corresponding set/
nj=nj-1；
if nj=0
jrank=layer+1；
Flayer+1=Flayer+1∪{xj}；
Layer+=1；/ * returns to Flayer*/
S3: improved crowding calculates
for fiAll objective function * of/* traversal/
for Fi/ * traverse level in all individual */
if fiIs first/* for first aim function, to initialize crowding */
id=0；
cal(fi, xi)；/ * calculating target function value */
desc(Fi, fi)；/ * descending arrangement */
istart=iend=∞；The both ends/* individual crowding be set as maximum */
for x1 to xsize-2/ * from the 2nd individual to second-to-last individual */
if x1=xi-1
id=(i-1)d；/ * same individual no longer calculate crowded angle value */
irank+=N；/ * label same individual */
else
S4: improved elite retention strategy
MergeG=baseG ∪ offsG；/ * parent and filial generation merge */
List[Fi]=NonDomination (mergeG)；The quick non-dominated ranking * of/*/
for List[Fi]/* traverse each level */
crowdingDistance(Fi)
sortbyDesc(Fi,d)；/ * is to FiAccording to crowding descending arrangement */
for Fi/ * by the individual of the level be incorporated to new population */
if irank≤ 2N/* if nonredundancy individual */
newG∪xi/ * be incorporated to new population */
else Ftemp∪xi
if newG>N
newG-(size-N)；/ * remove extra individual */
if newG<N
sortbyDesc(Ftemp)；/ * to interim level according to crowding sequence */
newG∪variation(N-size)；/ * the individual after variation be added to * in new population/
A variety of different dangerous material container stockpiling strategies can be obtained after the completion of calculating, do not conflicted between various strategies,
Strategy can be selected according to the actual situation.
It can be concluded that, the present invention is by analysis dangerous material container stockpiling regulation, to existing by above-described embodiment
Model for Multi-Objective Optimization improves, and proposes that new constraint optimizes, and it more close to the stockpiling of dangerous material stockyard
It asks, is solved using genetic algorithm, avoid locally optimal solution, improve operation efficiency, under conditions of guaranteeing dangerous cargo warehouse general safety,
Optimization stockpiling case position in the shortest possible time, provides a new point of penetration for harbour dangerous material container management, and subtract
It plans the container space time less, promotes the competition soft power and Hard Power at economic benefits and harbour.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (7)
1. a kind of harbour dangerous material container space distribution model and algorithm, which comprises the following steps:
S1: analytical integration is carried out to dangerous material and is summarized, Codes and Standards of dangerous material container during stockpiling are obtained, is such as divided
Class stockpiling, stockpiling height limitation etc., as one of the restrictive condition of multiple-objection optimization later；
S2: investigating the stockpiling situation of harbour dangerous cargo warehouse, and reservoir area and number of plies limitation can be stored up by obtaining all kinds of dangerous material
Specific requirement；
S3: building harbour dangerous material container stores up Model for Multi-Objective Optimization；
S4: setting multiple target improved adaptive GA-IAGA；
S5: harbour dangerous material container stockpiling Model for Multi-Objective Optimization is solved, obtains optimal dangerous material container case
Position stacked arrangement.
2. a kind of harbour dangerous material container space distribution model according to claim 1 and algorithm, it is characterised in that: institute
It states in S1 and the method that analytical integration summarizes is carried out to dangerous material, mainly by referring to such as GB 6944-2012, GB 12268-
2012, the relevant laws and regulations of JT 397-2007 and GB 18218-2009.
3. a kind of harbour dangerous material container space distribution model according to claim 1 and algorithm, it is characterised in that: institute
The building for stating model in S3 considers following hypothesis: not considering the difference of container size, is 20 ruler cabinets；Just for hazardous chemical
Product；It marches into the arena plan and container has been stored up it is known that including dangerous material container type, quality, marching into the arena the time in stockyard, if outlet
Case then corresponds to known to time for competiton, course line and port of destination.
4. a kind of harbour dangerous material container space distribution model according to claim 1 and algorithm, it is characterised in that: institute
That states model in S3 builds following set:
All container set, I={ 1,2,3 ..., Ni, and i, j ∈ I；
Container of showing up is total, Ni；
The set of field, K={ 1,2,3 ..., }；
The set of shellfish, B={ 01,03,05 ..., Nb}；
The set of row, R={ 1,2,3 ..., Nr}；
The set of layer, T={ 1,2,3 ..., Nt}；
Model builds following parameter in the S3:
SiExpression is put into specified place according to classification for premise, is 1 when meeting, and is 0 when not meeting；
Ei,jThe case first marched into the arena for premise is indicated in lower layer, is 1 when meeting, is 0 when not meeting；
Oi,jThe case first to appear on the scene for premise is indicated on upper layer, is 1 when meeting, is 0 when not meeting；
Wi,jIndicate same for premise voyage, weight case upper, be 1 when meeting, be 0 when not meeting；
Ob,r,tIndicate occupied case position；
Cb,r,tIndicate the transportation cost of different shellfishes；
Model builds following decision variable in the S3:
xi,b,r,t, otherwise it is 0 that container i, which is 1 when being placed to t layers of r column of b shellfish,；
Y, i and j are 1 when being assigned to two case positions, are otherwise 0；
zeIf the case that the stockpiling of container i and container j first march into the arena for premise in lower layer, is 1 when meeting, it is when not meeting
0；
zoIf the case that the stockpiling of container i and container j first appear on the scene for premise on upper layer, is 1 when meeting, it is when not meeting
0；
zwIf the case of the stockpiling of container i and container j voyage same for premise, weight is 1, when not meeting when meeting upper
It is 0；
qkThe actual mass unit for indicating dangerous cargo k stockpiling is ton, QkIndicate the critical mass list that dangerous cargo k can store up
Position is ton, type included in k ∈ subordinate list 1.
5. a kind of harbour dangerous material container space distribution model according to claim 1 and algorithm, it is characterised in that: institute
It states using dangerous material container storage standards as constraint condition in S3, is intended to be precondition to march into the arena, with safety, economy
Property for the purpose of carry out case bit optimization, construct the multi-objective Model of four targets: it is minimum on upper layer to be unsatisfactory for the case first to appear on the scene；No
The remote case in the same voyage of satisfaction rule, port of destination is upper minimum；The same voyage of rule, heavy case are unsatisfactory for upper minimum；Stockpiling height
It is minimum；Stockpiling case position is closer far from passageway, and transportation range is nearest.
6. a kind of harbour dangerous material container space distribution model according to claim 5 and algorithm, it is characterised in that: institute
The multi-objective Model for stating four targets is mainly made of objective function and constraint condition, and objective function is Min f={ fo, fw,
fh, fc, wherein
1. it is minimum on upper layer to be unsatisfactory for the case first to appear on the scene
fo=∑I, j, b, r, t1 < t2((y-zo)*(t2-t1)) 3-1
2. being unsatisfactory for the case of regular EXPORT CARTON, same to voyage, weight upper minimum
fw=∑I, j, b, r, t1 < t2((y-zw)*(t2-t1)) 3-2
3. minimizing stockpiling height.
fh=∑I, b, r, t(XI, b, r, t*f2) 3-3
4. transportation range is nearest.
fc=∑b(CB, r, t*∑B, r, t(XI, b, r, t)) 3-4
Constraint condition:
Wherein constraint condition 3-5 and 3-6 indicate to be placed on j as i in the following, and i ratio j is early when marching into the arena, ze1 is taken, is otherwise 0；3-7 and 3-8
Indicate decision variable y and XI, b, r, tBetween relationship, only when container i and j are respectively allocated to case position b, r, t1And b, r, t2
When, the value of decision variable y takes 1, is otherwise 0；3-9 indicates that chest cannot be placed vacantly；3-10 indicates that dangerous material are stored in and answers
The place of the stockpiling；3-11 indicates that each on-site will have baffle-box position.
7. a kind of harbour dangerous material container space distribution model according to claim 1 and algorithm, it is characterised in that: institute
The setting for stating multiple target improved adaptive GA-IAGA in S3 includes the following steps:
S1: initial population setting, in order to which the order of container and specific location more directly show, using real number
Mode encode, i.e., encoded with the position of container, a position of the chromosome of an individual indicates in stacking
Actual position, encoding scheme is as shown in the picture, indicates that the position of all possible storages after container is shown up has m, and preceding n is a
Indicate distribution position and the allocation order of container；
S2: quick non-dominated ranking, steps are as follows:
S201: after generating initial population, to each individual xiAll set a parameter niWith a set si, niIt is the individual quilt
The number for the solution that other individuals dominate, that is, the individual amount more outstanding than the individual；siIt is in the population dominated by the individual
Other groups of individuals, be in other words exactly group of individuals more worse than the individual；
S202: individual x is calculatediThe value in each optimization object function, that is, bring into formula 3-1,3-2,3-3,3-4, obtain
The functional value f of all targetso(xi), fw(xi), fh(xi), fc(xi)；
S203: in addition to xiOutside, other individuals x in population is traversedj, similarly, by each individual xjBring into respectively a formula 3-1,
In 3-2,3-3,3-4, the functional value f of the individual is obtainedo(xj), fw(xj), fh(xj), fc(xj), two individuals are compared about all
The value of objective function: if meeting fo(xi) < fo(xj), fw(xi) < fw(xj), fh(xi) < fh(xj), fc(xi) < fc(xj),
Then claim individual xiDominate individual xj, dominate set siIn to add individual xj；If meeting fo(xi) > fo(xj), fw(xi) > fw
(xj), fh(xi) > fh(xj), fc(xi) > fo(xj), then claim individual xiBy individual xjIt dominates, by domination parameter niIncrease by 1；
S204: individual x in population has been traversedjAfterwards, individual x is observediParameter ni: if it is equal to 0, illustrates individual xiIt should not be by population
In other individual dominate, it should non-dominant level F is added1；If it is greater than 0, illustrate individual xiNon-dominant layer cannot be incorporated into
Grade；
S205: returning in S202, continues to traverse other individuals, complete until traversing；
S206: non-dominant level F is generated according to S202-S2051, wherein individual xiCorresponding set siIn all individual
xj, to the parameter n of individual each in setj, n is executed to itj=nj- 1 operation, then judges njValue: if be equal to 0, explanation
Individual xjIt is currently no longer dominated by other individuals, level F can be added2；If being not equal to 0, explanation cannot be incorporated into level
F2；
S207: other individuals are looped through, until F1Traversal is complete；
S208: to non-dominant F2Do with the identical operation of S206, until next level be sky；S3: improved crowding meter
Calculation method, steps are as follows:
S301: the individual x in non-dominant level is traversedi, the functional value of some objective function is calculated, carries out descending according to functional value
Arrangement is set as maximum value to the crowding of the individual on sequence both sides, it is ensured that it is chosen to always；
S302: since the 2nd individual to second-to-last individual until, calculate individual xjCrowding before, first judge it
It is whether identical with individual before: if identical, no longer calculating individual xjCrowding, but directly gathered around
Crowded angle value is set as the crowded angle value of same individual, and by individual xjNon-dominant hierarchical value on add population scale, increase
The non-dominant level of the individual, reduces selected probability, realizes the label of redundancy individual, if it is not, then according to gathering around as follows
Degree formula is squeezed to be calculated:
S303: if there is multiple optimization object functions, return step S301 is until not traversed all optimization aims
Only；
S4: improved elite retention strategy is judged redundancy individual using label, and is incorporated into interim level, finally new
When the population scale deficiency of generation, it is taken out corresponding redundancy individual and is incorporated to new population, steps are as follows:
S401: parent and filial generation are merged into operation, population scale is made to be reached for 2N；
S402: quick non-dominated ranking is carried out according to S2 to the population after merging, determines each non-dominant level；
S403: to each non-dominant level, the crowding of individual is calculated according to S3；
S404: creation new population P respectively operates each non-dominant level, first to PiIn individual according to crowded angle value
Descending sort is carried out, each of these individual is then traversed, judges whether the non-dominant level of current individual is greater than 2N: if
It is to illustrate that this is a redundancy individual, refusal is incorporated into new population P, but puts it into interim level PtempIn；If
It is not the level then continued after traversal；If whole traversals is complete, the scale of new population is less than N, then in interim level
Individual is ranked up according to crowded angle value, and individual is successively then taken out out of interim level to be incorporated to new population according to sequence, and
Individual first carries out mutation operation to it before being incorporated to new population, until new population scale is N；
S5: genetic manipulation, steps are as follows:
S501: initial population is selected using wheel disc bet method is improved, steps are as follows:
S50101: from the population P for needing to carry out genetic manipulation individual according to crowded angle value carry out descending sort, i.e., first according to
Non-dominant level belonging to individual carries out ascending sort, and individual identical for non-dominant level carries out descending according still further to crowding
Sequence；
S50102: all crowdings are added summation, and calculate select probability:
Wherein, PiFor the probability that ith member in current iteration is selected, diFor the fitness value of ith member；
S50103: generating random number m, probability value be added, until accumulated probability is equal to or more than m；
S502: crossover process, mainly in such a way that any intersects, i.e., simple to intersect, steps are as follows:
S50201: cross-point locations are determined；
S50202: Duplication part is excluded；
S503: mutation operation, according to rand0, the 1 random number size generated judges whether to make a variation, if can make a variation,
Individual chromosome is encoded the genic value on some upper position to replace with the genic value on another position of the chromosome, is replaced
After the completion, the legitimacy of individual is examined, if it is illegal, then needs to adjust sequence to meet condition, until concept feasible.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910234955.6A CN110378644A (en) | 2019-03-27 | 2019-03-27 | A kind of harbour dangerous material container space distribution model and algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910234955.6A CN110378644A (en) | 2019-03-27 | 2019-03-27 | A kind of harbour dangerous material container space distribution model and algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110378644A true CN110378644A (en) | 2019-10-25 |
Family
ID=68248433
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910234955.6A Pending CN110378644A (en) | 2019-03-27 | 2019-03-27 | A kind of harbour dangerous material container space distribution model and algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110378644A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144806A (en) * | 2019-12-18 | 2020-05-12 | 青岛港国际股份有限公司 | Automatic loading method for dangerous goods container |
| CN111932203A (en) * | 2020-08-03 | 2020-11-13 | 上海海勃物流软件有限公司 | Automatic position selection and distribution method, terminal and medium for turning containers in container area |
| CN113537545A (en) * | 2020-04-15 | 2021-10-22 | 北京旷视机器人技术有限公司 | Order picking strategy determining method and device and electronic equipment |
| KR102325608B1 (en) * | 2020-11-18 | 2021-11-15 | 김태호 | Method for generating training data of artificial neural network for deep reinforcement training of pre-marshalling of container terminal, storage medium for recoding computer program that can the method |
| CN114219386A (en) * | 2021-11-12 | 2022-03-22 | 武汉众智鸿图科技有限公司 | Dangerous cargo container yard isolation checking method, system and storage medium |
| CN116070983A (en) * | 2023-03-28 | 2023-05-05 | 交通运输部水运科学研究所 | Dangerous cargo container safety dispatching method, system, equipment and storage medium |
| CN118396489A (en) * | 2024-07-01 | 2024-07-26 | 天津东方泰瑞科技有限公司 | Port dangerous cargo container yard site selection method and system based on improved NSGA-II |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070050115A1 (en) * | 2005-08-24 | 2007-03-01 | Rockwell Automation Technologies, Inc. | Model-based control for crane control and underway replenishment |
| CN101585453A (en) * | 2008-05-20 | 2009-11-25 | 上海海事大学 | Distribution Method for export container yard of container wharf |
| CN103246941A (en) * | 2013-05-21 | 2013-08-14 | 武汉大学 | Scheduling method for export container wharf pile-up space |
| CN103544586A (en) * | 2013-10-25 | 2014-01-29 | 东北大学 | Cargo allocation method for improving quay crane operation efficiency and vessel stability of containers |
| CN109019056A (en) * | 2018-08-28 | 2018-12-18 | 盐田国际集装箱码头有限公司 | A kind of Container Yard bilayer aerial conveyor vertical transport equipment dispatching method |
-
2019
- 2019-03-27 CN CN201910234955.6A patent/CN110378644A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070050115A1 (en) * | 2005-08-24 | 2007-03-01 | Rockwell Automation Technologies, Inc. | Model-based control for crane control and underway replenishment |
| CN101585453A (en) * | 2008-05-20 | 2009-11-25 | 上海海事大学 | Distribution Method for export container yard of container wharf |
| CN103246941A (en) * | 2013-05-21 | 2013-08-14 | 武汉大学 | Scheduling method for export container wharf pile-up space |
| CN103544586A (en) * | 2013-10-25 | 2014-01-29 | 东北大学 | Cargo allocation method for improving quay crane operation efficiency and vessel stability of containers |
| CN109019056A (en) * | 2018-08-28 | 2018-12-18 | 盐田国际集装箱码头有限公司 | A kind of Container Yard bilayer aerial conveyor vertical transport equipment dispatching method |
Non-Patent Citations (1)
| Title |
|---|
| 谢尘，等: "基于混堆模式的集装箱码头出口箱进场选位策略", 《上海海事大学学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144806A (en) * | 2019-12-18 | 2020-05-12 | 青岛港国际股份有限公司 | Automatic loading method for dangerous goods container |
| CN113537545A (en) * | 2020-04-15 | 2021-10-22 | 北京旷视机器人技术有限公司 | Order picking strategy determining method and device and electronic equipment |
| CN111932203A (en) * | 2020-08-03 | 2020-11-13 | 上海海勃物流软件有限公司 | Automatic position selection and distribution method, terminal and medium for turning containers in container area |
| CN111932203B (en) * | 2020-08-03 | 2021-06-22 | 上海海勃物流软件有限公司 | Automatic position selection and distribution method, terminal and medium for turning containers in container area |
| KR102325608B1 (en) * | 2020-11-18 | 2021-11-15 | 김태호 | Method for generating training data of artificial neural network for deep reinforcement training of pre-marshalling of container terminal, storage medium for recoding computer program that can the method |
| CN114219386A (en) * | 2021-11-12 | 2022-03-22 | 武汉众智鸿图科技有限公司 | Dangerous cargo container yard isolation checking method, system and storage medium |
| CN116070983A (en) * | 2023-03-28 | 2023-05-05 | 交通运输部水运科学研究所 | Dangerous cargo container safety dispatching method, system, equipment and storage medium |
| CN118396489A (en) * | 2024-07-01 | 2024-07-26 | 天津东方泰瑞科技有限公司 | Port dangerous cargo container yard site selection method and system based on improved NSGA-II |
| CN118396489B (en) * | 2024-07-01 | 2024-09-13 | 天津东方泰瑞科技有限公司 | Port dangerous cargo container yard site selection method and system based on improved NSGA-II |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110378644A (en) | A kind of harbour dangerous material container space distribution model and algorithm | |
| CN101585453B (en) | Distribution Method for export container yard of container wharf | |
| Van Den Berg | A literature survey on planning and control of warehousing systems | |
| Kozan et al. | Genetic algorithms to schedule container transfers at multimodal terminals | |
| Kim et al. | Heuristic algorithms for routing yard‐side equipment for minimizing loading times in container terminals | |
| Tanaka et al. | Solving real-world sized container pre-marshalling problems with an iterative deepening branch-and-bound algorithm | |
| CN103544586B (en) | A kind of stowage method improving container ship stability and bank bridge working performance | |
| WO2022252268A1 (en) | Optimized scheduling method for intelligent stereoscopic warehouse | |
| CN107480922A (en) | Both ends formula is unloaded goods bit allocation scheduling model method for building up with the double car operational modes of rail | |
| CN111861205B (en) | Automatic container wharf outlet box space distribution method | |
| CN103246941B (en) | Space scheduling method stored up by a kind of Containers For Export harbour | |
| CN109886478A (en) | A kind of slotting optimization method of finished wine automatic stereowarehouse | |
| CN107977756B (en) | Ternary tree planning calculation method for solving three-dimensional packing problem | |
| CN114841642B (en) | Auxiliary material warehouse entry cargo space distribution method based on eagle perch optimization | |
| CN110599000A (en) | Automated dock rollover evaluation method, box position distribution method and related device | |
| CN111598499B (en) | Order allocation strategy determination method and device and electronic equipment | |
| CN108861619A (en) | A kind of half mixes palletizing method, system and robot offline | |
| Ayachi et al. | A Genetic algorithm to solve the container storage space allocation problem | |
| CN111144806A (en) | Automatic loading method for dangerous goods container | |
| CN113570025A (en) | E-commerce storage center goods space distribution method based on discrete particle swarm algorithm | |
| Ozcan et al. | A reward-based algorithm for the stacking of outbound containers | |
| Razouk et al. | New approaches for solving the container stacking problem | |
| CN116451290A (en) | Container yard pre-dumping method | |
| CN110009138A (en) | A kind of three acidification work warehouse goods yard layout methods based on particle swarm optimization algorithm | |
| Lorenc et al. | Evaluation of the effectiveness of methods and criteria for product classification in the warehouse |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191025 |