CN108764555A - A kind of shared bicycle based on Hadoop parks a site selecting method - Google Patents
A kind of shared bicycle based on Hadoop parks a site selecting method Download PDFInfo
- Publication number
- CN108764555A CN108764555A CN201810493379.2A CN201810493379A CN108764555A CN 108764555 A CN108764555 A CN 108764555A CN 201810493379 A CN201810493379 A CN 201810493379A CN 108764555 A CN108764555 A CN 108764555A
- Authority
- CN
- China
- Prior art keywords
- bicycle
- demand
- shared bicycle
- population
- stop
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000005457 optimization Methods 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000013439 planning Methods 0.000 claims description 30
- 230000035772 mutation Effects 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 7
- 230000006978 adaptation Effects 0.000 claims description 5
- 230000010429 evolutionary process Effects 0.000 claims description 4
- 230000006798 recombination Effects 0.000 claims description 4
- 238000005215 recombination Methods 0.000 claims description 4
- 238000000205 computational method Methods 0.000 claims description 2
- 238000012217 deletion Methods 0.000 claims description 2
- 230000037430 deletion Effects 0.000 claims description 2
- 238000013178 mathematical model Methods 0.000 claims description 2
- 230000009467 reduction Effects 0.000 claims description 2
- 241001269238 Data Species 0.000 abstract description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 230000002068 genetic effect Effects 0.000 description 4
- 108090000623 proteins and genes Proteins 0.000 description 4
- 230000004075 alteration Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000010353 genetic engineering Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 206010038743 Restlessness Diseases 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000452 restraining effect Effects 0.000 description 1
- 238000009394 selective breeding Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of, and the shared bicycle based on Hadoop parks a site selecting method, including 1) the shared bicycle demand point prediction based on Distributed Cluster algorithm；2) the shared bicycle based on multiple-objection optimization parks a site selection model；3) the model solution algorithm based on NSGA-II algorithms；Algorithm is realized using Hadoop simultaneously, after initialization of population, the process evolved per a generation is completed with a MapReduce.The beneficial effects of the invention are as follows：The present invention proposes a kind of shared bicycle based on Hadoop and parks a site selecting method, and this method is estimated can to improve reasonability and accuracy that shared bicycle parks an addressing, makes the management more specification of shared bicycle；For with a large amount of shared bicycle trip datas of generation, the present invention establishes the Demand Forecast Model based on trip data and predicts shared bicycle demand point by Hadoop frames and clustering algorithm.
Description
Technical field
The present invention relates to technical field of computer information processing, and more specifically, it is related to a kind of shared based on Hadoop
Bicycle parks site selecting method a little.
Background technology
Recent years, along with the raising of social development and living standards of the people, change also occurs for people's consciousness of going on a journey,
Low-carbon trip becomes the theme of people's trip.The fusion product-of modern science and technology and public bicycles is shared bicycle and is asked therewith
Generation, and quickly occuping market core status.Shared bicycle compensates for the fixed point of public bicycles by means of returning the car, and cash pledge returns inconvenience
Etc. inadequate natural endowments, shared bicycle be then more in line with the traffic path of people, facilitate the trip of people conscientiously, the spy parked everywhere
Property cause a large number of users that shared bicycle has been selected to go on a journey, as the solution of " last one kilometer ", shared bicycle, which becomes,
The first choice that passerby rides instead of walk.
However while the rapid development of shared bicycle, a large amount of problems is also produced.It is closed since shared bicycle lacks
Removing the work office and planning of science activities, the unrest of the vehicle of generation stop leaving about, well damage, cannot clear up in time and cause part congestion in road etc.
Problem has all seriously affected people's lives.How reasonably to plan that parking for shared bicycle a little becomes to be even more important, if stopped
It is unreasonable to put an addressing, will lead to the problem of as traditional public bicycles, a large number of users is caused to abandon riding.It is building
Under the overall background in smart city and big data epoch, how just to seem to park a progress rational deployment and planning for sharing bicycle
It is very significant.
Patent CN201710764773.0 " a kind of shared bicycle park determination method and device a little " provide it is a kind of altogether
It enjoys bicycle and parks determination method and device a little, this method includes：Obtain the foot path data in predeterminable area；Based on the step
The position coordinates for the tracing point that row track includes cluster foot path using default clustering algorithm；According to each classification
Including foot path corresponding to real street path distribution situation, determine that shared bicycle is parked a little.The invention is determining altogether
It enjoys during bicycle parks a little, by foot path data, the use on which real street path of approach is reasonably determined out
Family has demand of riding, and shared bicycle is arranged on the street route and parks a little, and parking determination a little to shared bicycle has
Guiding significance, to serve more users with demand of riding so that shared bicycle resource allocation is more balanced.Specially
Profit 201710517669.1 " a kind of shared bicycle park determination method and device a little " provides a kind of shared bicycle and parks a little
Determination method and device, this method includes：According to it is corresponding with current slot have park the functional area information of demand,
The subregion with the function is determined from predeterminable area；Classified to determining subregion using default sorting algorithm；It will classification
The center of obtained each classification is determined as shared bicycle and parks a little.Determining that sharing bicycle parks mistake a little in the invention
Cheng Zhong associates time factor and the subregion with the demand of parking, and to current slot it is corresponding it is multiple have stop
The subregion for putting demand is classified, and is parked the center of each classification as shared bicycle a little, in this way can be effectively
Guidance is shared bicycle administrative staff and is scheduled to shared bicycle, to meet the user that current slot is located in predeterminable area
To sharing the demand of bicycle, reduces some region to the greatest extent and occur sharing bicycle the case where supply falls short of demand, promote user experience.But
Be these method and systems it is to realize to park prediction a little to sharing bicycle, does not make full use of the history of shared bicycle
Trip data can not judge the reasonability for parking a little, and parking of providing does not have a little yet and government department plans
Region is contacted, and parks accuracy a little and operability is not strong.
In conclusion the key of shared bicycle management is the reasonable distribution of user demand and bicycle quantity, it could fully
The maximum value for sharing bicycle is played, solves the problems, such as its generation.At present shared bicycle park an addressing there are accuracy compared with
It is low to cause relation between supply and demand unreasonable, cause resource allocation unreasonable, the problems such as managerial confusion.
Invention content
The purpose of the present invention is being directed to shared bicycle to park that an addressing is unreasonable to lead to shared bicycle managerial confusion, one is provided
Shared bicycle of the kind based on Hadoop parks a site selecting method.The present invention realizes by the following technical solutions：
This shared bicycle based on Hadoop parks a site selecting method, includes mainly three parts：It is poly- based on distribution
Shared bicycle demand point prediction, the shared bicycle based on multiple-objection optimization of class algorithm park a site selection model, are based on NSGA-II
The model solution algorithm of algorithm.
(1) the shared bicycle demand point prediction based on Distributed Cluster algorithm：The accuracy of demand point prediction, to shared single
Vehicle parks addressing a little and plays critical effect, and traditional requirement forecasting is based primarily upon experience and small-scale data statistics comes
It carries out, requirement forecasting is not accurate enough, causes the planning of lease point not reasonable.This patent is directed to shared bicycle and carries GPS positioning
System, the characteristics of having produced a large amount of real user trip datas, propose more rational demand point prediction model, to altogether
Bicycle demand point is enjoyed to be predicted.
(2) the shared bicycle based on multiple-objection optimization parks a site selection model：After above-mentioned demand point step, need
It asks a little and carries out addressing distribution between can planning a little, user's trip most short and shared bicycle of total distance is parked into total cost most
Small established as target parks a site selection model.
(3) the model solution algorithm based on NSGA-II algorithms：Above-mentioned model is a classical biobjective scheduling problem,
Two object functions cannot be optimal simultaneously in model, so the model, there are many feasible solutions, this patent is selected in more mesh
Model is solved on the basis of more mature NSGA-II algorithms in mark evolution algorithm, NSGA-II is calculated for the model
The problems such as method optimizes, and slow according to run time has carried out distributed improvement.
The overall structure of this method as shown in Figure 1, specific implementation steps are as follows：
Step 1: the shared bicycle demand point prediction based on Distributed Cluster algorithm
Shared bicycle release has produced mass users data so far, and the present invention is for these mass datas to shared single
The demand point of vehicle is predicted.These trip datas include time, bicycle number, bicycle type, GPS position information etc..Interception
Partial data is illustrated in fig. 2 shown below.By the way that the bicycle data at a certain moment, using clustering is carried out by way of clustering, formation is permitted
Mostly a certain range of demand region, we are using the cluster centre point in demand region as demand point, within the scope of demand region
Shared demand of the bicycle quantity as demand point.Demand point prediction model frame proposed by the present invention is as shown in figure 3, the model
Detailed process it is as follows：
1) actual conditions of demand point are directed to, two threshold values of Canopy are set, i.e. maximums of the T1 for demand point between away from
From T2 is the maximum magnitude of each demand point.
2) Canopy algorithms are executed, the number of demand point and the position of demand point are obtained.
3) demand point of generation is screened, by containing the less isolated point deletion of demand, obtains new data set.
4) using remaining demand point quantity as K values, demand point position as the initial cluster heart, by K-means algorithms into
Row iteration operation, finally obtains cluster result.
Step 2: the shared bicycle based on multiple-objection optimization parks a site selection model
It is exactly by the demand of demand point and each to be selected to advise for shared bicycle stop siteselecting planning problem is popular
It draws and optimizes the addressings of the quantity relations of distribution between stop, show that each shared bicycle demand point is distributed to each planning and parked
The bicycle quantity allotted of point.
The model is that bicycle parks a little total construction cost minimum and user's trip total distance is most short for optimization aim to share.
Concrete mathematical model is expressed as：
In formula：
I：Indicate the set { 1,2,3...i } of demand point；
J：Indicate that set { 1,2,3...j } a little is parked in planning；
ni：Indicate the bicycle demand of demand point i；
dij：Indicate the distance of demand point i to candidate planning stop j；
xij：Indicate that demand point i distributes to the bicycle quantity of candidate planning stop j；
cj：Indicate total bicycle quantity of stop j distribution after distributing；
M：Indicate the capital expenditure of each candidate planning stop；
c：Indicate the basic bicycle quantity of each candidate planning stop planning, often exceeding basic bicycle quantity one will increase
Add construction and administration fee Y；
yj：Indicate whether that building the candidate plans stop；
aj：Indicate that candidate planning stop exceeds the number of basic bicycle quantity；
Wherein, object function (1) makes the bicycle of demand point be minimized to the total distance of the candidate stop planned；Target letter
The total cost that number (2) makes stop need minimizes.Formula (3) indicates that the shared bicycle of demand point has been distributed on stop.Formula
(4) it is used for calculating the bicycle quantity of the stop after distribution.Formula (5) indicates if stop bicycle quantity is 0 after distributing, no
Build the planning stop.Formula (6) represents more than the number of the planning basic bicycle quantity of stop.
Step 3: the model solution algorithm based on NSGA-II algorithms
It is as follows for the algorithm solution procedure of the model：
Step 1：Read initial data, it is demand point set, facility candidate point set, the bicycle demand of each demand point, each
Distance, each candidate capital expenditure etc. of planning stop of the demand point to each candidate planning stop；
Step 2：By the way of matrix coder, population at individual is encoded, variable can be in value range, to population
Individual is initialized, and the population for including individual is generated.
Step 3：Calculate two target function values of each individual of population, root according to the fitness value of individual, to individual into
The quick non-dominated ranking of row.
Step 4：According to crowding computational methods, crowded angle value individual in population is calculated.
Step 5：According to improved adaptive crossover operator and mutation operator herein, find out the crossover probability of each individual with
Then mutation probability selects population, is intersected, mutation operation, new progeny population is generated.
Step 6：Using the mode of elitism strategy, merge parent and progeny population, forms big kind that population at individual number is 2N
Group.
Step 7：The calculating that quick non-dominated ranking and crowding are carried out to merging the population generated, finds out preferably N number of
Individual forms the parent population of a new generation.
Step 8：Repeat step 5.
Step 9：According to the adaptive adjustment to feasible solution and infeasible solution, judges whether to recombination and intersect.
Step 10：Step 6 is repeated, the progeny population of a new generation is obtained.
Step 11：Whether determining program evolutionary generation is more than greatest iteration number or to meet end condition, is then program knot
Beam, otherwise, t=t+1 goes to step 7 and continues to execute.
The algorithm performs flow is as shown in Figure 4.The optimal solution set that model can be found out by the algorithm, so as to stopping
Put a progress disjunctive programming.
Algorithm is realized using Hadoop simultaneously, after initialization of population, the process evolved per a generation is with one
MapReduce is completed.Wherein the Map stages are used for completing the calculating of individual adaptation degree, regard node subgroup number as key
Value, individual and its fitness are usually all relatively time-consuming in the process for completing these operations as value values, so using parallel
Operation；Reduce is responsible for the corresponding value values reduction of identical key values, then be directed to each node on subgroup into
The operations such as row selection, intersection, variation, can keep the relatively independent of subgroup evolutionary process.Since each node population is independent of each other,
By multiple Reduce nodes, also parallel form is used to carry out the evolutional operation of population.Parallelization flow chart such as Fig. 5 institutes
Show.
The beneficial effects of the invention are as follows:The present invention proposes a kind of shared bicycle based on Hadoop and parks an addressing side
Method, this method is estimated can to improve reasonability and accuracy that shared bicycle parks an addressing, make the management of shared bicycle more
Specification.For with a large amount of shared bicycle trip datas of generation, this patent is established the Demand Forecast Model based on trip data, is led to
Hadoop frames and clustering algorithm are crossed, shared bicycle demand point is predicted.It might not could be provided as demand point simultaneously
The problem of parking, establishes and an addressing mould is parked with the multiple target of the most short and total minimum target of construction cost of total distance of going on a journey
Type can calculate shared bicycle by the model and park position a little and the open ended shared bicycle scale in the position.Finally
Model is solved using improved NSGA-II, and calculating process is realized using Hadoop frames.To a certain extent
Solve the problems, such as that shared bicycle parks an addressing, making to park an addressing becomes more science, reasonable.
Description of the drawings
Fig. 1 is that the shared bicycle proposed by the present invention based on Hadoop parks a site selecting method overall construction drawing；
Fig. 2 is to share bicycle part in the present invention to ride datagram；
Fig. 3 is demand point prediction model frame of the present invention；
Fig. 4 is that the present invention improves NSGA-II algorithm solving model flow charts；
Fig. 5 is the algorithm solving model flow chart that the present invention realizes on Hadoop frames；
Fig. 6 is the Canopy-Kmeans parallelization implementation processes used to model in step 1 of the present invention；
Fig. 7 is the K-means Parallel Algorithm flows based on MapReduce in the present invention.
Specific implementation mode
The present invention is described further with reference to embodiment.The explanation of following embodiments is merely used to help understand this
Invention.It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, also
Can be with several improvements and modifications are made to the present invention, these improvement and modification also fall into the protection domain of the claims in the present invention
It is interior.
The shared bicycle based on Hadoop as an implementation parks a site selecting method, and specific implementation step is such as
Under：
Step 1: the shared bicycle demand point prediction based on Distributed Cluster algorithm
This patent is used realizes that Canopy-Kmeans algorithms seek Fig. 3 models based on Hadoop parallel methods
Solution.Under MapReduce frames, the method for solving of the model can be split as several subtasks by we, and detailed process is such as
Shown in Fig. 6, each dashed lined box includes independent MapReduce tasks in figure.The cooperation of Canopy, k-means algorithm makes
With can overcome uncertainty that artificial selection K values bring, avoid due to randomly selecting the initial cluster heart so as to cause part most
It is excellent and algorithm it is unstable, reduce the problems such as influence of the isolated point to cluster result, have very to the clustering performance of k-means algorithms
Big raising.
The GPS information data of collected shared bicycle certain time period are subjected to file consolidation, deposit HDFS texts first
In part system, the parallel execution of the Canopy algorithms of first stage is executed by Hadoop, will be exported in the form of a file, it is right
Cluster centre point in file carries out map visualization, filters isolated point, file after processing is stored in HDFS, under Hadoop one
Stage K-means algorithm carries out clustering processing, output treated file, i.e. demand point position, quantity and information of vehicles.As
The input data of phase III.
Step 2: the shared bicycle based on multiple-objection optimization parks a site selection model
This patent has done some assumed conditions, for improving model when foundation shares bicycle and parks point model to model
Feasibility.It is found by the analysis and research of this paper, shared bicycle stop location problem has following feature：
(1) it is the project in urban construction and development to share bicycle to park the ground of a selection, and the scale of such project is
It is long-term, additionally due to for sharing the higher fence design cost in bicycle stop addressing place, plan for land, daily
Great amount of cost is required in operation, that means that shared bicycle parks cost constraint in a site selection model and accounts for important one
Point.
(2) shared bicycle stop addressing area to be built is divided into multiple electronics according to its land status, geographical conditions
Fence area, since traveler selection pedestrian traffic mode is limited by distance factor, so each fence region
Bicycle parking capacity possesses certain upper limit, and construction and administration fee will be increased by being more than basic parking quantity.
(3) in addition to cost limits, the convenient degree of trip of traveler how is improved, and determines that sharing bicycle parking clicks
The key factor of location place quality, herein assume traveler using share bicycle go on a journey when, be bound to chosen distance its recently
Park a little.In order to fully meet the different demands for going out administrative staff, i.e. demand point reaches the distance parked a little for sharing bicycle
Most short, shared bicycle is parked position a little and could be accepted extensively by people, to solve the problems, such as that shared bicycle disorderly stops to leave about.For
The position of all shared bicycles in demand region is all considered as the position of demand point, fence by more preferable Optimized model herein
Center be considered as shared bicycle and park position a little.In order to make all bicycles in each demand region to shared bicycle as far as possible
The distance parked a little is all nearest, and model parks the total distance of a distance as excellent using the bicycle in all demand regions to planning
Change target, reduce all demand points to greatest extent to the greatest extent and parks distance a little.
Step 3: the model solution algorithm based on improved NSGA-II
1 coding mode
NSGA-II algorithms use real coding and binary coding mode, one-dimensional real coding and binary coding
It all cannot preferably reflect the various combined situations of population at individual in model.Set forth herein park a mould for shared bicycle
Type just has the mode that real number matrix encodes, to be encoded to population at individual.Specific form such as formula (3-1) indicates：
P in formulakFor k-th of individual in population；Xi,jFor corresponding i-th row of encoder matrix, jth column element, meaning
I demand point distributes to the bicycle quantity of j-th of stop；ViIndicate that i-th of demand point distributes to the distribution condition of stop；
RjIndicate distribution condition of j-th of stop from each demand point；
It is tied herein by by the way of matrix coder, can be good at reflecting in population at individual to NSGA-II algorithms
The allocation plan of fruit can make population at individual keep diversity in the operation for intersecting and making a variation, and avoid premature generation part
Convergence and precocious phenomenon.
2 crossover operators and mutation operator
Since population at individual uses the mode of real number matrix coding, the intersection of NSGA-II algorithms has been redesigned herein
Operator and mutation operator.Using fixed crossover operator and mutation operator in NSGA-II algorithms, due to crossover probability PcWith
PmFor fixed value, dynamic need of the population change procedure to these parameters is cannot be satisfied, according to these problems, gives new friendship
Pitch operator and mutation operator.
1) crossover operators：
Traditional crossover operator is generally by the way of single-point intersection and two-point crossover, in this way between population at individual
Gene exchange is inadequate, takes carry out crossover operation to a certain row in matrix herein.Two population at individual intersected are needed as follows：
It is then C by the individual that crossover operation generates1,C2, expression formula is as follows：
Wherein, i is the crosspoint generated at random, i between 1~N,
P1.rank individual P is represented1Non-dominated ranking level, P2.rank individual P is represented2Non-dominated ranking level.This
Text is by the way that the Pareto non-dominated ranking levels of the ginseng of crossover operator and each individual in population to be associated, and algorithm is in operation
Early period, due to large percentage of the small individual of Pareto non-dominated ranking values in offspring, the value of λ can be bigger, but with calculation
The continuous progress of method, individual tend to same Pareto leading surfaces, and the value of λ gradually tends to 0.5.Using this crossover operator strategy,
Preferable gene genetic in parent can be gone down, improve the diversity of population at individual.
2) mutation operators
For traditional mutation operator, it is typically employed in individual and a node is selected to carry out mutation operation.Due to using
The coding mode of real number matrix coding, therefore the mutation operation of a node cannot be used.Aggregation model and coding mode, herein
It is arranged into mutation operation using to a certain, it is as follows
P=[R1 P,R2 P,...,RN P] (3-7)
The individual P to make a variation is needed, is generated by variation：
Q=[R1 P,R2 P,...,Ri,...,RN P] (3-8)
RiFor the column data generated at random, the i-th original column data is replaced.
By above description it is known that the detailed process of crossover operator and mutation operator, in the parameter of genetic algorithm,
The key of performance of genetic algorithms is mainly crossover probability PcWith mutation probability PmSelection.Crossover probability PcBigger, new individual generates
Speed may be faster, if PmWhen excessive, and the possibility that hereditary pattern can be caused to be destroyed increases；PcIt is too small so as to search
Rope process is slow.For different optimization problems, need to test repeatedly to determine PcAnd Pm, each problem is adapted to it is difficult to find
Optimum value.Since NSGA-II algorithms are using fixed intersection and mutation probability, for this purpose, introducing M.Srinvivas herein
[44] et al. propose that a kind of Adaptive Genetic is calculated.
Thinking when individual adaptation degree is less than population average fitness in the strategy, it is possible to determine that the individual performance is bad,
It copes with it and assigns larger crossing-over rate and aberration rate, the individual with new model is promoted to generate；It is put down when individual adaptation degree is more than or equal to
When equal fitness, it is possible to determine that pattern gene possessed by the individual is more outstanding, copes with it and assigns smaller crossing-over rate and variation
Rate, to ensure that more excellent pattern gene is not destroyed in population.Its corresponding model is given below, formula (3-9) is crossing-over rate tune
Integral function, formula (3-10) are aberration rate Tuning function.
Wherein, PcFor individual intersection rate to be intersected, PmTo wait for variation individual aberration rate, fmaxFor in population at individual fitness most
Big value, favgFor population at individual average fitness, f', which is two, to be waited intersecting maximum adaptation degree in individual, and f is to wait for that variation individual adapts to
Degree, k1、k2For the parameter of crossing-over rate Tuning function, k3、k4For the parameter of crossing-over rate Tuning function.Under normal circumstances, k1=k2, k3
=k4。
The processing of 3 constrained optimizations improves
In practical application, their true optimal solution of the multi-objective optimization question of many belt restrainings is often possible to be present in about
Beam near border, these are located at the infeasible solution of restrained boundary its target function value and are often better than feasible zone inside points feasible solution
Target function value.So using the infeasible solution of these very advantageous, for improving to the close search speed of feasible zone.By
In model Existence restraint condition, population can be caused to generate infeasible solution during evolution, it is infeasible to kind in order to fully consider
The influence that group brings herein simultaneously accounts for feasible solution and infeasible solution, proposes then to choose every several generations of evolving more excellent
Set of feasible solution and infeasible disaggregation carry out genetic manipulation.
Throwback carries out recombination intersection to infeasible solution and feasible solution, judges to execute generation by a kind of adaptive strategy
Number.Because evolutionary process is evolved to feasible zone and optimal solution direction, feasible solution quantity can be more next in evolutionary process
It is more, if carrying out excessive genetic manipulation to feasible solution and infeasible solution again in later stage of evolution, calculation may be instead resulted in
Search performance of the method in feasible zone is affected, therefore herein using gradually reducing infeasible solution and feasible during evolution
The number that solution is directly intersected.Herein for this problem, set during feasible solution and infeasible solution throwback execute genetic manipulation
The algebraically that adaptive adjustment feasible solution is intersected with infeasible solution is set, that is to say, that just right when Evolution of Population algebraically is k
The two executes recombination and intersects:
In formula (3-11), T is the total evolutionary generation of population.It is right from formula it can be seen that as Evolution of Population algebraically increases
The operation of feasible solution and infeasible solution gradually decreases.
NSGA-II algorithms are encoded according to above-mentioned requirements, initial population scale N=100, initial crossover probability Pc
=0.8, mutation probability Pm=0.1, the maximum iteration of algorithm is max=100；The parallel of such as Fig. 5 is carried out using Hadoop
It executes, finally output can plan that stop Pareto optimality disaggregation is selected for policymaker.
Claims (4)
1. a kind of shared bicycle based on Hadoop parks a site selecting method, which is characterized in that include the following steps：
Step 1: the shared bicycle demand point prediction based on Distributed Cluster algorithm
Trip data includes time, bicycle number, bicycle type, GPS position information；It is adopted by the bicycle data to a certain moment
Clustering is carried out with the mode of cluster, many a certain range of demand regions are formed, by the cluster centre point in demand region
As demand point, demand of the shared bicycle quantity as demand point within the scope of demand region；
Step 2: the shared bicycle based on multiple-objection optimization parks a site selection model
Shared bicycle stop siteselecting planning problem be by the demand of demand point and it is each it is to be selected plan between stop into
The addressing of the row optimization quantity relations of distribution show that each shared bicycle demand point distributes to the bicycle that each planning is parked a little and distributes number
Amount；
The model is that bicycle parks a little total construction cost minimum and user's trip total distance is most short for optimization aim to share；Specifically
Mathematical model is expressed as：
In formula：
I：Indicate the set { 1,2,3...i } of demand point；
J：Indicate that set { 1,2,3...j } a little is parked in planning；
ni：Indicate the bicycle demand of demand point i；
dij：Indicate the distance of demand point i to candidate planning stop j；
xij：Indicate that demand point i distributes to the bicycle quantity of candidate planning stop j；
cj：Indicate total bicycle quantity of stop j distribution after distributing；
M：Indicate the capital expenditure of each candidate planning stop；
c：The basic bicycle quantity for indicating each candidate planning stop planning, often builds increase beyond basic bicycle quantity one
If with administration fee Y；
yj：Indicate whether that building the candidate plans stop；
aj：Indicate that candidate planning stop exceeds the number of basic bicycle quantity；
Wherein, object function (1) makes the bicycle of demand point be minimized to the total distance of the candidate stop planned；Object function
(2) total cost that stop needs is made to minimize；Formula (3) indicates that the shared bicycle of demand point has been distributed on stop；Formula (4)
For calculating the bicycle quantity of the stop after distribution；Formula (5) indicates, if stop bicycle quantity is 0 after distributing, not build
If the planning stop；Formula (6) represents more than the number of the planning basic bicycle quantity of stop；
Step 3: the model solution algorithm based on NSGA-II algorithms.
2. the shared bicycle according to claim 1 based on Hadoop parks a site selecting method, which is characterized in that step 1
Detailed process it is as follows：
1) actual conditions of demand point are directed to, two threshold values of Canopy, i.e. maximum distances of the T1 between demand point, T2 are set
For the maximum magnitude of each demand point；
2) Canopy algorithms are executed, the number of demand point and the position of demand point are obtained；
3) demand point of generation is screened, by containing the less isolated point deletion of demand, obtains new data set；
4) using remaining demand point quantity as K values, demand point position is changed as the initial cluster heart by K-means algorithms
For operation, cluster result is finally obtained.
3. the shared bicycle according to claim 1 based on Hadoop parks a site selecting method, which is characterized in that step 3
It is as follows for the algorithm solution procedure of the model：
Step 1：Read initial data, demand point set, facility candidate point set, the bicycle demand of each demand point, each demand
Point plans the distance of stop, the capital expenditure etc. of each candidate planning stop to each candidate；
Step 2：By the way of matrix coder, population at individual is encoded, variable can be in value range, to population at individual
It is initialized, generates the population for including individual；
Step 3：Two target function values of each individual of population are calculated, root carries out individual fast according to the fitness value of individual
Fast non-dominated ranking；
Step 4：According to crowding computational methods, crowded angle value individual in population is calculated；
Step 5：According to improved adaptive crossover operator and mutation operator herein, crossover probability and the variation of each individual are found out
Then probability selects population, is intersected, mutation operation, new progeny population is generated；
Step 6：Using the mode of elitism strategy, merge parent and progeny population, forms the big population that population at individual number is 2N；
Step 7：The calculating that quick non-dominated ranking and crowding are carried out to merging the population generated, finds out preferably individual,
Form the parent population of a new generation；
Step 8：Repeat step 5；
Step 9：According to the adaptive adjustment to feasible solution and infeasible solution, judges whether to recombination and intersect；
Step 10：Step 6 is repeated, the progeny population of a new generation is obtained；
Step 11：Whether determining program evolutionary generation is more than greatest iteration number or meets end condition, is then EP (end of program), no
Then, t=t+1 goes to step 7 and continues to execute.
4. the shared bicycle according to claim 3 based on Hadoop parks a site selecting method, which is characterized in that the step
Rapid San Tong method realizes algorithm using Hadoop, and after initialization of population, the process evolved per a generation is with one
MapReduce is completed；Wherein the Map stages are used for completing the calculating of individual adaptation degree, regard node subgroup number as key
Value, individual and its fitness are usually all relatively time-consuming in the process for completing these operations as value values, so using parallel
Operation；Reduce is responsible for the corresponding value values reduction of identical key values, then be directed to each node on subgroup into
The operations such as row selection, intersection, variation, can keep the relatively independent of subgroup evolutionary process；Since each node population is independent of each other,
By multiple Reduce nodes, also parallel form is used to carry out the evolutional operation of population.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810493379.2A CN108764555B (en) | 2018-05-22 | 2018-05-22 | Shared bicycle parking point site selection method based on Hadoop |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810493379.2A CN108764555B (en) | 2018-05-22 | 2018-05-22 | Shared bicycle parking point site selection method based on Hadoop |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108764555A true CN108764555A (en) | 2018-11-06 |
| CN108764555B CN108764555B (en) | 2021-08-31 |
Family
ID=64008547
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810493379.2A Active CN108764555B (en) | 2018-05-22 | 2018-05-22 | Shared bicycle parking point site selection method based on Hadoop |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108764555B (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902969A (en) * | 2019-03-13 | 2019-06-18 | 武汉大学 | A kind of shared bicycle release position planing method based on OD data |
| CN110059986A (en) * | 2019-05-08 | 2019-07-26 | 武汉大学 | A kind of shared bicycle dynamic put-on method and system |
| CN110147885A (en) * | 2019-05-07 | 2019-08-20 | 浙江大学城市学院 | A kind of shared bicycle of improved adaptive GA-IAGA parks point distribution method |
| CN110175704A (en) * | 2019-05-07 | 2019-08-27 | 南京师范大学 | A kind of shared vehicle specification parking area division methods |
| CN110458309A (en) * | 2019-06-29 | 2019-11-15 | 东南大学 | A kind of net based on practical road network environment about share-car bus station position method |
| CN111551958A (en) * | 2020-04-28 | 2020-08-18 | 北京踏歌智行科技有限公司 | Mining area unmanned high-precision map manufacturing method |
| CN111881939A (en) * | 2020-06-24 | 2020-11-03 | 东南大学 | Shared single-vehicle parking area layout method based on clustering algorithm |
| CN112131330A (en) * | 2020-09-16 | 2020-12-25 | 上海交通大学 | Site selection layout method for shared automobile operation area in free flow mode |
| CN112163788A (en) * | 2020-10-21 | 2021-01-01 | 深圳市规划国土发展研究中心 | Real-time data-based scheduling method for Internet pile-free single vehicle |
| CN112185014A (en) * | 2020-04-07 | 2021-01-05 | 江苏智途科技股份有限公司 | Method for judging rationality of parking points of shared bicycle |
| CN113095406A (en) * | 2021-04-14 | 2021-07-09 | 国能智慧科技发展(江苏)有限公司 | Electronic fence effective time period management and control method based on intelligent Internet of things |
| CN113095670A (en) * | 2021-04-08 | 2021-07-09 | 上海市城市建设设计研究总院（集团）有限公司 | Planning and site selection method for shared bicycle storage points |
| CN113920713A (en) * | 2021-09-29 | 2022-01-11 | 广州时空位置网科学技术研究院有限公司 | Urban traffic intelligent information management system based on Beidou positioning |
| CN114327859A (en) * | 2021-11-18 | 2022-04-12 | 西安电子科技大学 | Source model cluster selection method for cloud computing environment large-scale problem agent optimization |
| CN114386500A (en) * | 2021-12-31 | 2022-04-22 | 苏州市公安局 | Infectious disease sampling point selection method, device, equipment and storage medium |
| CN116665455A (en) * | 2023-07-18 | 2023-08-29 | 北京阿帕科蓝科技有限公司 | Vehicle station selection method, device and computer equipment |
| CN116702954A (en) * | 2023-05-16 | 2023-09-05 | 苏州大学 | Long-time dynamic address selection method and device for movable service facility and readable storage medium |
| CN112163788B (en) * | 2020-10-21 | 2024-05-31 | 深圳市规划国土发展研究中心 | Scheduling method of Internet pile-free bicycle based on real-time data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095266A (en) * | 2014-05-08 | 2015-11-25 | 中国科学院声学研究所 | Method and system for clustering optimization based on Canopy algorithm |
| CN107392239A (en) * | 2017-07-11 | 2017-11-24 | 南京邮电大学 | A kind of K Means algorithm optimization methods based on Spark computation models |
| CN107463620A (en) * | 2017-07-05 | 2017-12-12 | 洛川闰土农牧科技有限责任公司 | A kind of elevator accident early-warning and predicting system based on data mining |
| CN107871184A (en) * | 2017-11-16 | 2018-04-03 | 南京邮电大学 | A kind of site selecting method of the electric automobile charging station of facing area electrically-charging equipment |
| CN108038575A (en) * | 2017-12-20 | 2018-05-15 | 广西大学 | Waypoint location planing method based on modified NSGA II |
-
2018
- 2018-05-22 CN CN201810493379.2A patent/CN108764555B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095266A (en) * | 2014-05-08 | 2015-11-25 | 中国科学院声学研究所 | Method and system for clustering optimization based on Canopy algorithm |
| CN107463620A (en) * | 2017-07-05 | 2017-12-12 | 洛川闰土农牧科技有限责任公司 | A kind of elevator accident early-warning and predicting system based on data mining |
| CN107392239A (en) * | 2017-07-11 | 2017-11-24 | 南京邮电大学 | A kind of K Means algorithm optimization methods based on Spark computation models |
| CN107871184A (en) * | 2017-11-16 | 2018-04-03 | 南京邮电大学 | A kind of site selecting method of the electric automobile charging station of facing area electrically-charging equipment |
| CN108038575A (en) * | 2017-12-20 | 2018-05-15 | 广西大学 | Waypoint location planing method based on modified NSGA II |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902969A (en) * | 2019-03-13 | 2019-06-18 | 武汉大学 | A kind of shared bicycle release position planing method based on OD data |
| CN110147885A (en) * | 2019-05-07 | 2019-08-20 | 浙江大学城市学院 | A kind of shared bicycle of improved adaptive GA-IAGA parks point distribution method |
| CN110175704A (en) * | 2019-05-07 | 2019-08-27 | 南京师范大学 | A kind of shared vehicle specification parking area division methods |
| CN110175704B (en) * | 2019-05-07 | 2022-12-27 | 南京师范大学 | Method for dividing standard parking area of shared vehicles |
| CN110059986A (en) * | 2019-05-08 | 2019-07-26 | 武汉大学 | A kind of shared bicycle dynamic put-on method and system |
| CN110059986B (en) * | 2019-05-08 | 2021-02-19 | 武汉大学 | Dynamic releasing method and system for shared bicycle |
| CN110458309A (en) * | 2019-06-29 | 2019-11-15 | 东南大学 | A kind of net based on practical road network environment about share-car bus station position method |
| CN110458309B (en) * | 2019-06-29 | 2023-07-11 | 东南大学 | Network about splicing station point location method based on actual road network environment |
| CN112185014A (en) * | 2020-04-07 | 2021-01-05 | 江苏智途科技股份有限公司 | Method for judging rationality of parking points of shared bicycle |
| CN111551958B (en) * | 2020-04-28 | 2022-04-01 | 北京踏歌智行科技有限公司 | Mining area unmanned high-precision map manufacturing method |
| CN111551958A (en) * | 2020-04-28 | 2020-08-18 | 北京踏歌智行科技有限公司 | Mining area unmanned high-precision map manufacturing method |
| CN111881939A (en) * | 2020-06-24 | 2020-11-03 | 东南大学 | Shared single-vehicle parking area layout method based on clustering algorithm |
| CN111881939B (en) * | 2020-06-24 | 2021-03-09 | 东南大学 | Shared single-vehicle parking area layout method based on clustering algorithm |
| CN112131330B (en) * | 2020-09-16 | 2024-01-26 | 上海交通大学 | Method for selecting and laying out operation area of shared automobile in free flow mode |
| CN112131330A (en) * | 2020-09-16 | 2020-12-25 | 上海交通大学 | Site selection layout method for shared automobile operation area in free flow mode |
| CN112163788A (en) * | 2020-10-21 | 2021-01-01 | 深圳市规划国土发展研究中心 | Real-time data-based scheduling method for Internet pile-free single vehicle |
| CN112163788B (en) * | 2020-10-21 | 2024-05-31 | 深圳市规划国土发展研究中心 | Scheduling method of Internet pile-free bicycle based on real-time data |
| CN113095670A (en) * | 2021-04-08 | 2021-07-09 | 上海市城市建设设计研究总院（集团）有限公司 | Planning and site selection method for shared bicycle storage points |
| CN113095406A (en) * | 2021-04-14 | 2021-07-09 | 国能智慧科技发展(江苏)有限公司 | Electronic fence effective time period management and control method based on intelligent Internet of things |
| CN113920713A (en) * | 2021-09-29 | 2022-01-11 | 广州时空位置网科学技术研究院有限公司 | Urban traffic intelligent information management system based on Beidou positioning |
| CN113920713B (en) * | 2021-09-29 | 2024-01-23 | 广州时空位置网科学技术研究院有限公司 | Urban traffic intelligent information management system based on Beidou positioning |
| CN114327859A (en) * | 2021-11-18 | 2022-04-12 | 西安电子科技大学 | Source model cluster selection method for cloud computing environment large-scale problem agent optimization |
| CN114327859B (en) * | 2021-11-18 | 2024-06-07 | 西安电子科技大学 | Source model clustering selection method for large-scale problem agent optimization of cloud computing environment |
| CN114386500A (en) * | 2021-12-31 | 2022-04-22 | 苏州市公安局 | Infectious disease sampling point selection method, device, equipment and storage medium |
| CN116702954A (en) * | 2023-05-16 | 2023-09-05 | 苏州大学 | Long-time dynamic address selection method and device for movable service facility and readable storage medium |
| CN116665455A (en) * | 2023-07-18 | 2023-08-29 | 北京阿帕科蓝科技有限公司 | Vehicle station selection method, device and computer equipment |
| CN116665455B (en) * | 2023-07-18 | 2023-12-01 | 北京阿帕科蓝科技有限公司 | Vehicle station selection method, device and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108764555B (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764555A (en) | A kind of shared bicycle based on Hadoop parks a site selecting method | |
| Boyacı et al. | An integrated optimization-simulation framework for vehicle and personnel relocations of electric carsharing systems with reservations | |
| CN100444210C (en) | Mixed controlling method of single dot signal controlling crossing | |
| Sabar et al. | A memetic algorithm for real world multi-intersection traffic signal optimisation problems | |
| CN105809279B (en) | A kind of water resources optimal operation method based on multi-target quantum shuffled frog leaping algorithm | |
| CN108364467A (en) | A kind of traffic information prediction technique based on modified decision Tree algorithms | |
| CN106570595A (en) | Subway station position selection method and system based on space big data | |
| CN109377048A (en) | A kind of comprehensive traffic network hub node selection method | |
| CN110288210A (en) | Project level road infrastructure comprehensive maintenance decision-making technique and system | |
| Shen et al. | Optimized public parking location modelling for green intelligent transportation system using genetic algorithms | |
| CN112230675B (en) | Unmanned aerial vehicle task allocation method considering operation environment and performance in collaborative search and rescue | |
| CN103050016B (en) | Hybrid recommendation-based traffic signal control scheme real-time selection method | |
| CN104077634B (en) | active-reactive type dynamic project scheduling method based on multi-objective optimization | |
| CN110046213A (en) | A kind of electric power selection method for taking path distortion correction and scissors crossing correction into account | |
| CN106781464A (en) | A kind of congestion in road situation method of testing | |
| CN107506433A (en) | Urban development space general layout Scene Simulation system | |
| CN107832957A (en) | A kind of public lease bicycle dispatching method based on the handling of polymorphic type bicycle | |
| CN114626718A (en) | Agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing | |
| CN113053122A (en) | WMGIRL algorithm-based regional flow distribution prediction method in variable traffic control scheme | |
| CN116029470A (en) | Public transport route planning method based on crowd sensing | |
| CN110222133A (en) | A kind of multiclass land use simulating and predicting method and device considering landscape index | |
| Lee et al. | Equity based land‐use and transportation problem | |
| CN116797126A (en) | Rural terminal logistics site selection-path planning method based on double-layer planning | |
| Yu et al. | Optimization of urban bus operation frequency under common route condition with rail transit | |
| Wang et al. | Multi-reservoir system operation theory and practice |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220719 Address after: 310015 No. 51, Huzhou street, Hangzhou, Zhejiang Patentee after: Zhejiang University City College Address before: 310015 No. 50 Huzhou Street, Hangzhou City, Zhejiang Province Patentee before: Zhejiang University City College |