CN104504229B - A kind of intelligent public transportation dispatching method based on hybrid metaheuristics - Google Patents

A kind of intelligent public transportation dispatching method based on hybrid metaheuristics Download PDF

Info

Publication number
CN104504229B
CN104504229B CN201410481840.4A CN201410481840A CN104504229B CN 104504229 B CN104504229 B CN 104504229B CN 201410481840 A CN201410481840 A CN 201410481840A CN 104504229 B CN104504229 B CN 104504229B
Authority
CN
China
Prior art keywords
mrow
msub
fitness
individual
new
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.)
Active
Application number
CN201410481840.4A
Other languages
Chinese (zh)
Other versions
CN104504229A (en
Inventor
郑宁
陈涛
徐海涛
林菲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Dianzi University
Original Assignee
Hangzhou Dianzi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hangzhou Dianzi University filed Critical Hangzhou Dianzi University
Priority to CN201410481840.4A priority Critical patent/CN104504229B/en
Publication of CN104504229A publication Critical patent/CN104504229A/en
Application granted granted Critical
Publication of CN104504229B publication Critical patent/CN104504229B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a kind of intelligent public transportation dispatching method based on hybrid metaheuristics.Simulated annealing and genetic algorithm are combined together by the present invention, and add elite retention strategy and fitness stretch function.The maximum individual of fitness in every generation population is directly remained into the next generation, avoids it from being intersected and is destroyed with mutation operation.Fitness stretch function algorithm initial stage, cut down individual between difference, so as to increase the diversity of population, avoid genetic algorithm from being absorbed in locally optimal solution；Difference between the later stage of algorithm, increase individual, so as to increase the selected probability of excellent individual, accelerate convergence rate.Arithmetic speed of the present invention is fast, the operation plan after being optimized, the stand-by period of passenger can be greatly reduced within a short period of time under given time of departure frequency condition；Departure frequency is adapted dynamically, departure frequency is met the changing rule of passenger flow total amount；The departure interval is adapted dynamically, the stand-by period of passenger is greatly reduced.

Description

A kind of intelligent public transportation dispatching method based on hybrid metaheuristics
Technical field
The invention belongs to city intelligent public transportation system technical field, is related to a kind of intelligence based on hybrid metaheuristics Energy bus dispatching method, the method that particularly a kind of departure plan to bus rapid transit is scheduled.
Background technology
With the fast development in city, urban population is continuously increased, and the quantity of private car increases also with a large amount of, causes Traffic congestion and problem of environmental pollution are increasingly serious.The effect of urban public tranlport system is given full play to, these can be alleviated and asked Topic.But public transport also has the problem of passenger waiting time is longer and passenger satisfaction is relatively low at present.How effectively to solve The problem of existing, it is the key for increasing public transport attraction.Bus dispatching is the center of the daily operation activity of enterprises of public transport, it Directly influence operation cost and passenger satisfaction.The bus dispatching scheme for meeting passenger flow rule can be according to the change of the volume of the flow of passengers The departure interval is adjusted, the specific aim of bus service is strengthened, reduces the Waiting time of passenger, improve transit quality of service, Increase the attraction of public transport.
The purpose of bus dispatching is exactly in the case where meeting passenger's trip requirements, saves operation cost as far as possible.The two Conflicting requirement causes this to be a multi-objective optimization question.Meanwhile bus dispatching will by enterprises of public transport's operation cost, Many constraints such as Fleet size.Under the conditions of how meeting volume of the flow of passengers demand and constraint at the same time, find suitable method and exist Bus dispatching scheme is determined in the rational time, is the key for realizing intelligent bus scheduling.Bus dispatching is divided into static scheduling With two parts of dynamic dispatching, static scheduling is primarily referred to as working out the departure time-table of every circuit, and dynamic dispatching is mainly completed Existing departure time-table is adjusted when there are the emergency cases such as vehicle, passenger flow.In daily operation, static scheduling is It is main, supplemented by dynamic dispatching.The invention mainly relates to how using improved hybrid metaheuristics to solve the quiet of bus rapid transit State scheduling problem.
At present, the research both at home and abroad in this field has a lot, but the concrete condition of each urban mass-transit system is different, There is no a kind of more general method that can combine history operation data progress bus dispatching.There are many people in these researchs Using heuritic approaches such as genetic algorithms, or their innovatory algorithm is used for the optimization of bus dispatching.Genetic algorithm can be The optimal solution or approximate optimal solution of bus dispatching problem are found in the rational time, this is also that many researchers are solved using it Certainly the reason for bus dispatching.But easy Premature Convergence be present in genetic algorithm, the shortcomings such as efficiency is low.
The content of the invention
The invention aims to overcome the shortcomings of genetic algorithm itself, and a kind of novel method is proposed to solve The conflict of interest between passenger and bus operation enterprise.Method of the present invention combines simulated annealing and genetic algorithm Together, elite retention strategy and fitness stretch function are added and.The maximum individual of fitness in every generation population is straight Connect and remain into the next generation, avoid it from being intersected and destroyed with mutation operation.Fitness stretch function is in the initial stage of algorithm, reduction Difference between individual, so as to increase the diversity of population, genetic algorithm is avoided to be absorbed in locally optimal solution；In the later stage rank of algorithm Section, the difference between increase is individual, so as to increase the selected probability of excellent individual, accelerate convergence rate.Simulated annealing energy Enough increase the local search ability of genetic algorithm, so as to accelerate the convergence rate of genetic algorithm.
The method of the present invention comprises the following steps that：
Step (1) reads the record of swiping the card of passenger, and the number that statistics is ridden in total number of persons and each period I daily, obtains Take daily weather history situation and festivals or holidays situation；Using hierarchy clustering method, cluster point is carried out to the historical data of acquisition Analysis；The total number of persons by bus among one day, number of passengers, weather condition and the festivals or holidays situation in different time sections are combined into One vector, is then normalized operation to the vector, and cluster operation, root are carried out using the Ward methods in system clustering algorithm Each class another characteristic is extracted according to the result of cluster；
Described record of swiping the card includes get on the bus charge time, website of getting on the bus, get off charge time and get-off stop；
The historical data of described acquisition includes record of swiping the card, weather history situation and the festivals or holidays situation of passenger；
Weather forecast information and festivals or holidays situation of the step (2) according to second day, are matched from the cluster result of step 1 One class, and a vector is extracted as predicted value from such；
According to predicted value, with reference to the desired load factor of enterprises of public transport, both step (3) be divided by, and obtains total hair of second day Vehicle shift time；
Step (4) generates N number of vector at random, and vectorial dimension is equal with order of classes or grades at school of always dispatching a car；Each component represents corresponding class At the secondary time of departure, the time of departure, setting one-component was equal to 0, and last component is equal to last bus in units of minute The number of minutes between the time of departure and the first bus time of departure；Component in vector by order arrangement from small to large, this it is N number of to The set P of amount composition initial solution0, and iterations g is set as 0；Wherein N is even number；
Step (5) establishes the mathematical modeling of bus dispatching, most short for goal-setting fitness letter with the stand-by period of passenger Number, calculates the fitness of each initial solution, is then solved by hybrid metaheuristics；
5-1. carries out fitness stretched operation using fitness stretch function, and original adaptation is replaced with the value after stretching Degree；
5-2. is according to roulette selection strategy from set PgMiddle selection any two solution, is carried out according to the crossover probability of setting Crossover operation, that is, a crossover location is randomly choosed, exchange the part before and after two solution crosspoints, after obtaining two intersections Solution；Then solution carries out simulated annealing operation after intersecting to two：The fitness of the solution after intersecting is calculated, if fitness increases, Receive new solution, otherwise receive new solution with current acceptance probability；So as to obtain two new solutions；
For 5-3. according to the mutation probability of setting, each component of two new explanations to being obtained in step 5-2 enters row variation Operation, i.e., one size of generation is located at the natural number between former and later two components at random, replaces original value, obtains two changes Solution after different；Then simulated annealing operation is carried out：The fitness of the solution after variation is calculated, if fitness increases, is received new Solution, otherwise receive new solution with current acceptance probability；So as to obtain two new solutions, and two new solutions of acquisition are put into Disaggregation Pg+1；
5-4. repeat step 5-2 and 5-3, until disaggregation Pg+1The number of middle solution is equal with N；
Step (6) updates iterations g=g+1, if the iterations G having been maxed outmax, then disaggregation P is exportedgIn Fitness highest solution, the departure time-table after as optimizing；Otherwise, step 5-1 is gone to；
Described iterations GmaxFor positive integer.
Establishing for bus dispatching mathematical modeling described in step 5 is specific as follows：
(1) the bus dispatching mathematical modeling described in has following precondition：
After BRT vehicles stop, waiting Passengen is all got on the bus, in the absence of trapping phenomena；
BRT all fronts outside car by the way of charging, in the absence of the influence inserted coins or swiped the card to time of vehicle operation；
BRT completely uses identical vehicle model, and amount of seats is identical with maximum passenger carrying capacity；
Do not consider circuit matches somebody with somebody car, it is believed that vehicle is enough；
BRT vehicles are dispatched a car by schedule time list；
BRT vehicles order and sequence consensus of dispatching a car on road, in the absence of phenomenon of overtaking other vehicles；
Least unit using minute as scheduling；
(2) variable and its implication used in the bus dispatching mathematical modeling described in is as follows：
M is the number of dispatching a car in whole dispatching cycle；
N is website quantity total on circuit assigned direction；
tiFor the time of departure of ith car in a dispatching cycle, in units of minute, i=1,2 ..., m；
rjThe arrival rate changed over time for j-th of website on circuit assigned direction, unit behaviour/minute, j=1, 2,...,n；
T is the total waiting time of passenger in dispatching cycle；
Then
The cost of bus operation is divided into fixed cost and variable cost, is not present between bus dispatching and fixed cost direct Relation；If bus operation income is R in a dispatching cycle, P uses (member) for riding fee per capita, and L is that assigned direction circuit is total Length (km), C are the variable cost (member/km) of public transit vehicle, then the income of public transport company is that total income subtracts total become This, it is specific as follows：
If μ is passenger waiting time weight coefficient, ν is public transport company's income weight coefficient；According to secondary penalty method, bus dispatching The object function of Optimized model is：
Minz=μ × T- ν × R
If NmaxFor maximum appearance of vehicle amount, ρ is expectation load factor, then Prescribed Properties I：
In order to ensure rule reach passenger and it is random reach passenger and can wait until to wait in the short period of time, if HminAnd HmaxRespectively public transport company require the minimum and maximum departure interval, then the departure interval should meet following constraints Ⅱ：
Hmin≤ti-ti-1≤HmaxI=2,3 ..., m
Simultaneously to avoid causing the non-continuous event dispatched a car, if τ is the maximum departure interval difference that public transport company allows, then Constraints III：
|(ti+1-ti)-(ti-ti-1) |≤τ i=2,3 ..., m-1
Solution caused by being ensured by penalty function method does not violate the constraints of scheduling problem；Description based on more than, adjust Object function min f (X) form of degree problem is as follows：
Wherein min f (X) be add penalty function after target function value, ω1、ω2、ω3、ω4Respectively constraints I, II minimum departure interval of constraints, II maximum departure interval of constraints, penalty coefficient corresponding to constraints III；Scheduling is asked The solution X of topic be length be m vector, each component xiThe time of departure of the ith spacing first bus in dispatching cycle is represented, to divide Clock is unit；
In order to ensure that each individual fitness is all higher than 0, and also to which convenient use roulette selection strategy, to mesh Scalar functions, which enter line translation and obtain the final form of fitness function, is：
Described simulated annealing can strengthen the local search ability of genetic algorithm, be completed every time in genetic algorithm After crossover operation and mutation operation, compare former and later two individual fitness, carry out simulated annealing operation；Simulated annealing In need to set initial temperature T0, the calculation formula of Current Temperatures is：
T*=T0×σg-1
Wherein σ represents the speed that temperature reduces, and its value is 0 ＜ σ ＜ 1, and its value is bigger, and temperature is reduced slower, and value is got over Small, temperature reduces faster；G is the number of algorithm current iteration；When the fitness of new caused individual reduces, receive new individual The probability of body is：
Wherein F (Xnew)、F(Xold) it is respectively new individual and former individual fitness；
Simultaneously in order to ensure most to have in each population outstanding individual physical efficiency smoothly to enter the new individual of generation of future generation, in mould In the blending algorithm for intending annealing and genetic algorithm, elite retention strategy is added；After often producing new generation population, compare newer generation The fitness value of optimized individual in population and previous generation populations；If the fitness of the optimized individual of population of new generation is less than upper one The optimized individual in generation, then replace the individual that fitness is minimum in a new generation with previous generation optimized individual；Otherwise it is directly entered down An iteration.
The form of stretch function described in step 5-1 is：
Wherein F (Xi) represent individual XiFitness, F (Xi) ' it is fitness after stretching, T* refers in simulated annealing Current temperature, N represent quantity individual in population, and λ represents drawing coefficient；
In order to carry out fitness stretching, fitness criteria is carried out to individual all in population before the stretching, Order
Wherein fitness refers to the fitness of individual, and fitness' refers to the fitness after standardization, and max fitness, which refer to, to be worked as The fitness of optimized individual in preceding population.
The basic step of hybrid metaheuristics described in step 5 is as follows：
1) sets the value of following parameter：Population Size N, chromosome length Lc, crossover probability Pc, mutation probability Pm, it is maximum Evolutionary generation Gmax, initial temperature T0, annealing speed σ, drawing coefficient λ；
2) initializes population P0, that is, randomly generate N number of feasible be deconstructed into and initialize population P0；Calculate individual in initial population Fitness, be standardized, to fitness carry out stretched operation；Iterations g is set as 0；
3) carries out roulette selection according to the fitness after stretching, from population PgIn randomly select two individuals；
4) intersects and simulated annealing operates；Using single-point Crossover Strategy, by selected two individual p1、p2By probability Pc Carry out crossover operation and produce two new individual c1、c2；If F (ci) ＞ F (pi), then receive individual c1, otherwise with probability exp ((F(ci)-F(pi))/T*) receive new individual；
5) variations and simulated annealing operation；To new caused individual c1、c2Mutation operation is carried out by turn, if after variation Individual c1' fitness increase, then receive variation, otherwise with probability exp ((F (ci')-F(ci))/T*) receive new individual；
6) new caused two individuals are added new population P byg+1In, if Pg+1Middle individual amount is less than N, then goes to step It is rapid 3), otherwise carry out in next step；
7) calculates the fitness of each individual in new population, and is standardized operation；
8) carries out fitness stretching to the individual in new population；
9) implements elite retention strategy, and original seed group is replaced with new population；
10) coolings operation；
11) updates iterations g=g+1, if reaching maximum iteration Gmax, then population P is exportedgIn it is optimal Solution, otherwise goes to step 3).
The present invention has the beneficial effect that：
Fitness drawing process is added in Genetic Simulated Annealing Algorithm by the present invention, is improved former algorithm and is easily received too early The shortcomings that holding back.Compared with former algorithm, the algorithm after improvement is more preferable to the objective function optimization effect of bus dispatching problem.The present invention The method being related to optimizes the operation cost of bus operation enterprise, and passenger waiting time optimization is divided to enter two different stages OK.Bus operation cost is divided into fixed cost and variable cost, and bus dispatching is main and variable cost is closely related, and becomes This is mainly determined by order of classes or grades at school of dispatching a car.The optimization of operation cost is that the volume of the flow of passengers is reasonably predicted based on big data, Ran Houjie Close the vehicle load factor that operation enterprise it is expected to reach and calculate total order of classes or grades at school of dispatching a car.Cost Optimization Approach of the present invention can be with " my god " level on realize people the more dispatches a car, and people dispatches a car less effect.The optimization of passenger waiting time is by using melting The Genetic Simulated Annealing Algorithm of suitable response drawing process optimizes for the object function established, after finally giving optimization Bus dispatching plan.
Specifically realize following target：
It can be extracted each by the record of swiping the card of passenger according to whether being that festivals or holidays and weather condition carry out cluster analysis Class another characteristic.
The volume of the flow of passengers can be relatively defined according to the festivals or holidays and weather condition of second day on the basis of historical data True prediction.
The volume of the flow of passengers that can be obtained according to prediction, rational order of classes or grades at school of dispatching a car is calculated, the order of classes or grades at school that makes to dispatch a car is according to the volume of the flow of passengers It is how many to be adjusted correspondingly.
Can be under the conditions of given order of classes or grades at school of dispatching a car, according to different points of the volume of the flow of passengers in different time sections among one day Cloth situation, is adjusted to the departure interval, realizes that people the more dispatches a car within the different periods, the purpose that people dispatches a car less.But Maximum departure interval and minimum departure interval need to meet some requirements.
In summary, implementation result of the present invention is as follows：
(1) arithmetic speed is fast, can be within a short period of time under given time of departure frequency condition, the scheduling after being optimized Plan, is greatly reduced the stand-by period of passenger.(2) can be according to festivals or holidays and weather condition, dynamic adjusts departure frequency, Departure frequency is set to meet the changing rule of passenger flow total amount.(3) can be moved according to the change of the different time sections volume of the flow of passengers among one day State adjusts the departure interval, and the stand-by period of passenger is greatly reduced.
Brief description of the drawings
Fig. 1 flow charts of the present invention.
Fig. 2 history optimized individual fitness change curves.
Fig. 3 history population average fitness change curves.
Fig. 4 hybrid metaheuristics history optimized individuals fitness changes the contrast with other two kinds of algorithms.
Fig. 5 fitness stretched operation flow charts.
Fig. 6 crossover operation flow charts.
Fig. 7 mutation operation flow charts.
Specific embodiment
The invention will be further described with reference to the accompanying drawings and examples.
A kind of intelligent public transportation dispatching method based on hybrid metaheuristics, specifically comprises the following steps：
Step (1) reads the record of swiping the card of passenger, the number that statistics is ridden in total number of persons and each period I daily.From Daily weather history situation and festivals or holidays situation are captured in 2345 weather forecast websites and perpetual calendar.Use hierarchical clustering side Method, cluster analysis is carried out to the historical data of acquisition；I.e. by the total number of persons by bus among one day, the rider in different time sections Number, weather condition and festivals or holidays situation are combined into a vector, and operation then is normalized to the vector, uses Hierarchical Clustering Ward methods in algorithm carry out cluster operation, and each class another characteristic is extracted according to the result of cluster.After these operations, We can be accurately predicted the volume of the flow of passengers according to the concrete condition of scheduling day.
Record of swiping the card includes get on the bus charge time, website of getting on the bus, get off charge time and get-off stop.
The historical data of acquisition includes record of swiping the card, weather history situation and the festivals or holidays situation of passenger.
Weather forecast information and festivals or holidays situation of the step (2) according to second day, are matched from the cluster result of step 1 One class, and a vector is extracted as predicted value from such；
According to predicted value, with reference to the desired load factor of enterprises of public transport, both step (3) be divided by, and obtains total hair of second day Vehicle shift time；
Step (4) generates N number of vector at random, and vectorial dimension is equal with order of classes or grades at school of always dispatching a car.Each component represents corresponding class At the secondary time of departure, the time of departure, setting one-component was equal to 0, and last component is equal to last bus in units of minute The number of minutes between the time of departure and the first bus time of departure；Component in vector by order arrangement from small to large, this it is N number of to The set P of amount composition initial solution0, and iterations g is set as 0；Wherein N is even number；
Step (5) establishes the mathematical modeling of bus dispatching, most short for goal-setting fitness letter with the stand-by period of passenger Number, calculates the fitness of each initial solution, is then solved by hybrid metaheuristics.
5-1. carries out fitness stretched operation using fitness stretch function, and original adaptation is replaced with the value after stretching Degree；
5-2. is according to roulette selection strategy from set PgMiddle selection any two solution, is carried out according to the crossover probability of setting Crossover operation, that is, a crossover location is randomly choosed, exchange the part before and after two solution crosspoints, after obtaining two intersections Solution；Then the solution after intersecting to two carries out simulated annealing operation：The fitness of the solution after intersecting is calculated, if fitness increases Greatly, receive new solution, otherwise receive new solution with current acceptance probability；So as to obtain two new solutions；
For 5-3. according to the mutation probability of setting, each component of two new explanations to being obtained in step 5-2 enters row variation Operation, i.e., one size of generation is located at the natural number between former and later two components at random, replaces original value, obtains two changes Solution after different.Then simulated annealing operation is carried out：The fitness of the solution after variation is calculated, if fitness increases, is received new Solution, otherwise receive new solution with current acceptance probability；So as to obtain two new solutions, and two new solutions of acquisition are put into Disaggregation Pg+1；
5-4. repeat step 5-2 and 5-3, until disaggregation Pg+1The number of middle solution is equal with N；
Step (6) updates iterations g=g+1, if the iterations G having been maxed outmax, then disaggregation P is exportedgIn Fitness highest solution, the departure time-table after as optimizing.Otherwise, step 5-1 is gone to.
Described iterations GmaxFor positive integer.
Establishing for bus dispatching mathematical modeling described in step 5 is specific as follows：
(1) the bus dispatching mathematical modeling described in has following precondition：
After BRT vehicles stop, waiting Passengen is all got on the bus, in the absence of trapping phenomena.
BRT all fronts outside car by the way of charging, in the absence of the influence inserted coins or swiped the card to time of vehicle operation.
BRT completely uses identical vehicle model, and amount of seats is identical with maximum passenger carrying capacity.
Do not consider circuit matches somebody with somebody car, it is believed that vehicle is enough.
BRT vehicles are dispatched a car by schedule time list.
BRT vehicles order and sequence consensus of dispatching a car on road, in the absence of phenomenon of overtaking other vehicles.
Least unit using minute as scheduling.
(2) variable and its implication used in the bus dispatching mathematical modeling described in is as follows：
M is the number of dispatching a car in whole dispatching cycle；
N is website quantity total on circuit assigned direction；
tiFor the time of departure of ith car in a dispatching cycle, in units of minute, i=1,2 ..., m；
rjThe arrival rate changed over time for j-th of website on circuit assigned direction, unit behaviour/minute, j=1, 2,...,n；
T is the total waiting time of passenger in dispatching cycle.
Then
The cost of bus operation is divided into fixed cost and variable cost, is not present between bus dispatching and fixed cost direct Relation, therefore only consider variable cost here.If bus operation income is R in a dispatching cycle, P uses for riding fee per capita (member), L are assigned direction total line length (km), and C is the variable cost (member/km) of public transit vehicle.The income of public transport company is Total income subtracts total variable cost, therefore has
If μ is passenger waiting time weight coefficient, ν is public transport company's income weight coefficient.According to secondary penalty method, bus dispatching The object function of Optimized model is
Minz=μ × T- ν × R
In order to make full use of public transport resources, public transport company requires that vehicle load factor is higher than expectation load factor.If Nmax For maximum appearance of vehicle amount, ρ is expectation load factor, then Prescribed Properties I：
In order to ensure that rule reaches passenger and random reach passenger and can wait until to wait in the short period of time.Assuming that HminAnd HmaxRespectively public transport company require the minimum and maximum departure interval, then the departure interval should meet following constraints Ⅱ：
Hmin≤ti-ti-1≤HmaxI=2,3 ..., m
Meanwhile the difference of the departure interval of two neighboring order of classes or grades at school vehicle should not be too big, avoids causing the non-continuous event dispatched a car. Assuming that τ is the maximum departure interval difference that public transport company allows, then constraints III：
|(ti+1-ti)-(ti-ti-1) |≤τ i=2,3 ..., m-1
Solution caused by being ensured by penalty function method does not violate the constraints of scheduling problem；Description based on more than, adjust Object function min f (X) form of degree problem is as follows：
Wherein min f (X) be add penalty function after target function value, ω1、ω2、ω3、ω4Respectively constraints I, II minimum departure interval of constraints, II maximum departure interval of constraints, penalty coefficient corresponding to constraints III.Scheduling is asked The solution X of topic be length be m vector, each component xiThe time of departure of the ith spacing first bus in dispatching cycle is represented, to divide Clock is unit.
The object function of bus dispatching is minimization problem, in order to ensure each individual (among each corresponding disaggregation of individual A solution) fitness be all higher than 0, use roulette selection strategy and also to convenient, line translation entered to object function The final form for obtaining fitness function is：
Simulated annealing can strengthen the local search ability of genetic algorithm, complete to intersect behaviour every time in genetic algorithm Make, with after mutation operation, to compare former and later two individual fitness, carry out simulated annealing operation.Needed in simulated annealing Set initial temperature T0, the calculation formula of Current Temperatures is：
T*=T0×σg-1
Wherein σ represents the speed that temperature reduces, and its value is 0 ＜ σ ＜ 1, and its value is bigger, and temperature is reduced slower, and value is got over Small, temperature reduces faster；G is the number of algorithm current iteration.When the fitness of new caused individual reduces, receive new individual The probability of body is：
Wherein F (Xnew)、F(Xold) it is respectively new individual and former individual fitness.
Simultaneously in order to ensure most to have in each population outstanding individual physical efficiency smoothly to enter the new individual of generation of future generation, in mould In the blending algorithm for intending annealing and genetic algorithm, elite retention strategy is added.After often producing new generation population, compare newer generation The fitness value of optimized individual in population and previous generation populations.If the fitness of the optimized individual of population of new generation is less than upper one The optimized individual in generation, then replace the individual that fitness is minimum in a new generation with previous generation optimized individual.Otherwise it is directly entered down An iteration.
The form of stretch function is in step 5-1：
Wherein F (Xi) represent individual XiFitness, F (Xi) ' it is fitness after stretching, T* refers in simulated annealing Current temperature, N represent quantity individual in population, and λ represents drawing coefficient.
Embodiment 1
Effect of the stretch function in Genetic Simulated Annealing Algorithm is exemplified below.Work as T*=5000, λ=200 are false It is respectively 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 provided with 10 individual fitness, it is individual before and after stretching The comparing result of the selected probability of body is as shown in table 1.Work as T*=50, λ=200, it is assumed that there are 10 individual fitness difference For 0.71,0.72,0.73,0.74,0.75,0.76,0.77,0.78,0.79,0.80, the selected probability of individual before and after stretching Comparing result it is as shown in table 2.No. represents individual numbering in form, and fitness represents the fitness of individual, P1Represent stretching The preceding selected probability of individual, P2Represent the selected probability of individual after stretching.From comparing result, hence it is evident that can obtain following Conclusion：When temperature is higher, after stretched, the difference between individual reduces；When the temperature is low, it is stretched, individual Between difference become big.
The contrast of the selected probability of individual before and after being stretched during table 1T*=5000
No. 1 2 3 4 5 6 7 8 9 10
fitness 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
P1 0.0182 0.0364 0.0545 0.0727 0.0909 0.1091 0.1273 0.1455 0.1636 0.1818
P2 0.0982 0.0986 0.0990 0.0994 0.0998 0.1002 0.1006 0.1010 0.1014 0.1018
The contrast of the selected probability of individual before and after being stretched during table 2T*=50
No. 1 2 3 4 5 6 7 8 9 10
fitness 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.80
P1 0.0940 0.0954 0.0967 0.0980 0.0993 0.1007 0.1020 0.1033 0.1046 0.1060
P2 0.0830 0.0864 0.0899 0.0936 0.0974 0.1013 0.1055 0.1098 0.1143 0.1189
The digital scope expressed by basic data type that can be provided due to programming language is limited, in order to enter Row fitness stretches, and to carry out fitness criteria, order to individual all in population before the stretching
Wherein fitness refers to the fitness of individual, and fitness ' refers to the fitness after standardization, and max fitness, which refer to, to be worked as The fitness of optimized individual in preceding population.
Finally solved using the hybrid metaheuristics after improvement.Fig. 1 gives the flow chart of the algorithm, from figure It can be seen that improved hybrid metaheuristics mainly contain selection, intersection, variation, simulated annealing, elite reservation, standardization With fitness stretched operation, in solution procedure these operations act on colony P one by oneg, then produce population P of new generationg+1.Figure 2nd, 3,4 be history optimized individual fitness change curve, history population average fitness change song during algorithm performs respectively The comparison diagram of line and the algorithm and other two kinds of algorithms.
The basic step of hybrid metaheuristics described in step 5 is as follows：
12) sets the value of following parameter：Population Size N, chromosome length Lc, crossover probability Pc, mutation probability Pm, it is maximum Evolutionary generation Gmax, initial temperature T0, annealing speed σ, drawing coefficient λ.
13) initializes population P0, that is, randomly generate N number of feasible be deconstructed into and initialize population P0.Calculate individual in initial population The fitness of body, is standardized, and stretched operation is carried out to fitness.The flow chart for the stretched operation that Fig. 5 is provided.Set iteration Number g is 0.
14) carries out roulette selection according to the fitness after stretching, from population PgIn randomly select two individuals.
15) intersects and simulated annealing operates.Using single-point Crossover Strategy, by selected two individual p1、p2By probability PcCarry out crossover operation and produce two new individual c1、c2.If F (ci) ＞ F (pi), then receive individual c1, otherwise with probability exp ((F(ci)-F(pi))/T*) receive new individual.What Fig. 6 was provided is the flow chart of the step.
16) variations and simulated annealing operation.To new caused individual c1、c2Mutation operation is carried out by turn, if after variation Individual c1' fitness increase, then receive variation, otherwise with probability exp ((F (ci')-F(ci))/T*) receive new individual.Fig. 7 What is provided is the flow chart of the step.
17) new caused two individuals are added new population P byg+1In, if Pg+1Middle individual amount is less than N, then goes to step It is rapid 3), otherwise carry out in next step.
18) calculates the fitness of each individual in new population, and is standardized operation.
19) carries out fitness stretching to the individual in new population.
20) implements elite retention strategy, and original seed group is replaced with new population.
21) coolings operation.
22) updates iterations g=g+1, if reaching maximum iteration Gmax, then population P is exportedgIn it is optimal Solution, otherwise goes to step 3).

Claims (5)

  1. A kind of 1. intelligent public transportation dispatching method based on hybrid metaheuristics, it is characterised in that comprise the following steps：
    Step (1) reads the record of swiping the card of passenger, and the number in total number of persons and each period I, acquisition are every by bus daily for statistics It weather history situation and festivals or holidays situation；Using hierarchy clustering method, cluster analysis is carried out to the historical data of acquisition；I.e. Total number of persons by bus among one day, number of passengers, weather condition and the festivals or holidays situation in different time sections are combined into one Vector, operation then is normalized to the vector, cluster operation is carried out using the Ward methods in system clustering algorithm, according to poly- The result of class extracts each class another characteristic；
    Described record of swiping the card includes get on the bus charge time, website of getting on the bus, get off charge time and get-off stop；
    The historical data of described acquisition includes record of swiping the card, weather history situation and the festivals or holidays situation of passenger；
    Weather forecast information and festivals or holidays situation of the step (2) according to second day, match one from the cluster result of step 1 Class, and a vector is extracted as predicted value from such；
    According to predicted value, with reference to the desired load factor of enterprises of public transport, both step (3) be divided by, and obtains the always class of dispatching a car of second day It is secondary；
    Step (4) generates N number of vector at random, and vectorial dimension is equal with order of classes or grades at school of always dispatching a car；Each component represents corresponding order of classes or grades at school At the time of departure, the time of departure, setting one-component was equal to 0, and last component is dispatched a car equal to last bus in units of minute The number of minutes between time and the first bus time of departure；Component in vector is by order arrangement from small to large, this N number of Vector Groups Into the set P of initial solution0, and iterations g is set as 0；Wherein N is even number；
    Step (5) establishes the mathematical modeling of bus dispatching, most short for goal-setting fitness function with the stand-by period of passenger, meter The fitness of each initial solution is calculated, is then solved by hybrid metaheuristics；
    5-1. carries out fitness stretched operation using fitness stretch function, and original fitness is replaced with the value after stretching；
    5-2. is according to roulette selection strategy from set PgMiddle selection any two solution, is intersected according to the crossover probability of setting Operation, that is, a crossover location is randomly choosed, exchange the part behind two solution crosspoints, obtain the solution after two intersections；Then Solution carries out simulated annealing operation after intersecting to two：The fitness of the solution after intersecting is calculated, if fitness increases, is received new Solution, otherwise receive new solution with current acceptance probability；So as to obtain two new solutions；
    5-3. carries out mutation operation according to the mutation probability of setting to each component of two new explanations obtained in step 5-2, I.e. one size of generation is located at the natural number between former and later two components at random, replaces original value, after obtaining two variations Solution；Then simulated annealing operation is carried out：The fitness of the solution after variation is calculated, if fitness increases, receives new solution, it is no Then receive new solution with current acceptance probability；So as to obtain two new solutions, and two new solutions of acquisition are put into disaggregation Pg+1；
    5-4. judges disaggregation Pg+1Whether the number of middle solution is equal with N；If equal, implement elite retention strategy, that is, work as Pg+1 In the fitness of optimal solution compare PgIn fitness hour of optimal solution use disaggregation PgThe maximum solution of middle fitness replaces Pg+1The minimum solution of middle fitness；Otherwise, repeat step 5-2 and 5-3, until disaggregation Pg+1The number of middle solution is equal with N；
    Step (6) updates iterations g=g+1, if the iterations G having been maxed outmax, then disaggregation P is exportedgMiddle fitness Highest solution, the departure time-table after as optimizing；Otherwise, step 5-1 is gone to；
    Described iterations GmaxFor positive integer.
  2. A kind of 2. intelligent public transportation dispatching method based on hybrid metaheuristics as claimed in claim 1, it is characterised in that step Establishing for bus dispatching mathematical modeling described in rapid 5 is specific as follows：
    (1) the bus dispatching mathematical modeling described in has following precondition：
    After BRT vehicles stop, waiting Passengen is all got on the bus, in the absence of trapping phenomena；
    BRT all fronts outside car by the way of charging, in the absence of the influence inserted coins or swiped the card to time of vehicle operation；
    BRT completely uses identical vehicle model, and amount of seats is identical with maximum passenger carrying capacity；
    Do not consider circuit matches somebody with somebody car, it is believed that vehicle is enough；
    BRT vehicles are dispatched a car by schedule time list；
    BRT vehicles order and sequence consensus of dispatching a car on road, in the absence of phenomenon of overtaking other vehicles；
    Least unit using minute as scheduling；
    (2) variable and its implication used in the bus dispatching mathematical modeling described in is as follows：
    M is the number of dispatching a car in whole dispatching cycle；
    N is website quantity total on circuit assigned direction；
    tiFor the time of departure of ith car in a dispatching cycle, in units of minute, i=1,2 ..., m；
    rjThe arrival rate changed over time for j-th of website on circuit assigned direction, unit behaviour/minute, j=1,2 ..., n；
    T is the total waiting time of passenger in dispatching cycle；
    Then
    <mrow> <mi>T</mi> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>r</mi> <mi>j</mi> </msub> <mo>&amp;times;</mo> <mfrac> <msup> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mn>2</mn> </msup> <mn>2</mn> </mfrac> </mrow>
    The cost of bus operation is divided into fixed cost and variable cost, is not present between bus dispatching and fixed cost and directly closes System；If bus operation income is R in a dispatching cycle, P uses for riding fee per capita, and unit is member, and L is that assigned direction circuit is total Length, unit km, C are the variable cost of public transit vehicle, and unit is member/km, then the income of public transport company subtracts for total income Total variable cost, it is specific as follows：
    <mrow> <mi>R</mi> <mo>=</mo> <mi>P</mi> <mo>&amp;times;</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>r</mi> <mi>j</mi> </msub> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mi>C</mi> <mo>&amp;times;</mo> <mi>L</mi> <mo>&amp;times;</mo> <mi>m</mi> </mrow>
    If μ is passenger waiting time weight coefficient, ν is public transport company's income weight coefficient；According to secondary penalty method, bus dispatching optimization The object function of model is：
    Minz=μ × T- ν × R
    If NmaxFor maximum appearance of vehicle amount, ρ is expectation load factor, then Prescribed Properties I：
    <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>r</mi> <mi>j</mi> </msub> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>&amp;GreaterEqual;</mo> <mi>&amp;rho;</mi> <mo>&amp;times;</mo> <msub> <mi>N</mi> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> </msub> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>,</mo> <mo>...</mo> <mo>,</mo> <mi>m</mi> </mrow>
    In order to ensure rule reach passenger and it is random reach passenger and can wait until to wait in the short period of time, if HminWith HmaxRespectively public transport company require the minimum and maximum departure interval, then the departure interval should meet following constraints II：
    Hmin≤ti-ti-1≤HmaxI=2,3 ..., m
    Simultaneously to avoid causing the non-continuous event dispatched a car, if τ is the maximum departure interval difference that public transport company allows, then constrain Condition III：
    |(ti+1-ti)-(ti-ti-1) |≤τ i=2,3 ..., m-1
    Solution caused by being ensured by penalty function method does not violate the constraints of scheduling problem；Description based on more than, scheduling are asked Object function min f (X) form of topic is as follows：
    <mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <mi>min</mi> <mi> </mi> <mi>f</mi> <mrow> <mo>(</mo> <mi>X</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>min</mi> <mi> </mi> <mi>z</mi> <mo>+</mo> <msub> <mi>&amp;omega;</mi> <mn>1</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>2</mn> </mrow> <mi>m</mi> </munderover> <mo>{</mo> <mi>max</mi> <mo>{</mo> <mn>0</mn> <mo>,</mo> <mi>&amp;rho;</mi> <mo>&amp;times;</mo> <msub> <mi>N</mi> <mi>max</mi> </msub> <mo>-</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>r</mi> <mi>j</mi> </msub> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>}</mo> <mo>}</mo> <mo>+</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>&amp;omega;</mi> <mn>2</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>2</mn> </mrow> <mi>m</mi> </munderover> <mo>{</mo> <mi>max</mi> <mo>{</mo> <mn>0</mn> <mo>,</mo> <msub> <mi>H</mi> <mi>min</mi> </msub> <mo>-</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>}</mo> <mo>}</mo> <mo>+</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>&amp;omega;</mi> <mn>3</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>2</mn> </mrow> <mi>m</mi> </munderover> <mo>{</mo> <mi>max</mi> <mo>{</mo> <mn>0</mn> <mo>,</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>H</mi> <mi>max</mi> </msub> <mo>}</mo> <mo>}</mo> <mo>+</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>&amp;omega;</mi> <mn>4</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>2</mn> </mrow> <mi>m</mi> </munderover> <mo>{</mo> <mi>max</mi> <mo>{</mo> <mn>0</mn> <mo>,</mo> <mo>|</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> </mrow> </msub> <mo>-</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mrow> <mo>(</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mo>|</mo> <mi>&amp;tau;</mi> <mo>}</mo> <mo>}</mo> </mrow> </mtd> </mtr> </mtable> </mfenced>
    Wherein min f (X) be add penalty function after target function value, ω1、ω2、ω3、ω4Respectively constraints I, constraint II minimum departure interval of condition, II maximum departure interval of constraints, penalty coefficient corresponding to constraints III；Scheduling problem Solution X be length be m vector, each component xiRepresent the time of departure of the ith spacing first bus in dispatching cycle, using minute as Unit；
    In order to ensure that each individual fitness is all higher than 0, and also to which convenient use roulette selection strategy, to target letter Number, which enters line translation and obtains the final form of fitness function, is：
  3. A kind of 3. intelligent public transportation dispatching method based on hybrid metaheuristics as claimed in claim 1, it is characterised in that institute The simulated annealing stated can strengthen the local search ability of genetic algorithm, in genetic algorithm every time complete crossover operation and After mutation operation, compare former and later two individual fitness, carry out simulated annealing operation；Need to set in simulated annealing Initial temperature T0, the calculation formula of Current Temperatures is：
    T*=T0×σg-1
    Wherein σ represents the speed that temperature reduces, and its value is 0 ＜ σ ＜ 1, and its value is bigger, and temperature reduces slower, and value is smaller, temperature Degree reduces faster；G is the number of algorithm current iteration；When the fitness of new caused individual reduces, receive new individual Probability is：
    <mrow> <mi>P</mi> <mo>*</mo> <mo>=</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mi>F</mi> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mrow> <mi>n</mi> <mi>e</mi> <mi>w</mi> </mrow> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mrow> <mi>o</mi> <mi>l</mi> <mi>d</mi> </mrow> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mi>T</mi> <mo>*</mo> </mrow> </mfrac> <mo>)</mo> </mrow> </mrow>
    Wherein F (Xnew)、F(Xold) it is respectively new individual and former individual fitness；
    Simultaneously in order to ensure most to have in each population outstanding individual physical efficiency smoothly to enter the new individual of generation of future generation, moved back in simulation In fire and the blending algorithm of genetic algorithm, elite retention strategy is added；After often producing new generation population, than newer generation population With the fitness value of optimized individual in previous generation populations；If the fitness of the optimized individual of population of new generation is less than previous generation's Optimized individual, then replace the individual that fitness is minimum in a new generation with previous generation optimized individual；Otherwise it is directly entered next time Iteration.
  4. A kind of 4. intelligent public transportation dispatching method based on hybrid metaheuristics as claimed in claim 1, it is characterised in that step Suddenly the form of the stretch function described in 5-1 is：
    <mrow> <mi>F</mi> <msup> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <msup> <mi>e</mi> <mrow> <mi>&amp;lambda;</mi> <mo>&amp;times;</mo> <mi>F</mi> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>/</mo> <mi>T</mi> <mo>*</mo> </mrow> </msup> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msup> <mi>e</mi> <mrow> <mi>&amp;lambda;</mi> <mo>&amp;times;</mo> <mi>F</mi> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>/</mo> <mi>T</mi> <mo>*</mo> </mrow> </msup> </mrow> </mfrac> </mrow>
    Wherein F (Xi) represent individual XiFitness, F (Xi) ' it is fitness after stretching, T* refers in simulated annealing current Temperature, N represents quantity individual in population, and λ represents drawing coefficient；
    In order to carry out fitness stretching, fitness criteria, order are carried out to individual all in population before the stretching
    <mrow> <msup> <mi>fitness</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <mrow> <mi>f</mi> <mi>i</mi> <mi>t</mi> <mi>n</mi> <mi>e</mi> <mi>s</mi> <mi>s</mi> </mrow> <mrow> <mi>max</mi> <mi> </mi> <mi>f</mi> <mi>i</mi> <mi>t</mi> <mi>n</mi> <mi>e</mi> <mi>s</mi> <mi>s</mi> </mrow> </mfrac> </mrow>
    Wherein fitness refers to the fitness of individual, and fitness' refers to the fitness after standardization, and maxfitness refers to current population The fitness of middle optimized individual.
  5. A kind of 5. intelligent public transportation dispatching method based on hybrid metaheuristics as claimed in claim 1, it is characterised in that step The basic step of hybrid metaheuristics described in rapid 5 is as follows：
    1) sets the value of following parameter：Population Size N, chromosome length Lc, crossover probability Pc, mutation probability Pm, maximum evolution Algebraically Gmax, initial temperature T0, annealing speed σ, drawing coefficient λ；
    2) initializes population P0, that is, randomly generate N number of feasible be deconstructed into and initialize population P0；Calculate individual in initial population fit Response, it is standardized, stretched operation is carried out to fitness；Iterations g is set as 0；
    3) carries out roulette selection according to the fitness after stretching, from population PgIn randomly select two individuals；
    4) intersects and simulated annealing operates；Using single-point Crossover Strategy, by selected two individual p1、p2By probability PcCarry out Crossover operation produces two new individual c1、c2；If F (ci) ＞ F (pi), then receive individual c1, otherwise with probability exp ((F (ci)-F(pi))/T*) receive new individual；
    5) variations and simulated annealing operation；To new caused individual c1、c2Mutation operation is carried out by turn, if the individual after variation c1' fitness increase, then receive variation, otherwise with probability exp ((F (ci')-F(ci))/T*) receive new individual；
    6) new caused two individuals are added new population P byg+1In, if Pg+1Middle individual amount is less than N, then goes to step 3), Otherwise carry out in next step；
    7) calculates the fitness of each individual in new population, and is standardized operation；
    8) carries out fitness stretching to the individual in new population；
    9) implements elite retention strategy, and original seed group is replaced with new population；
    10) coolings operation；
    11) updates iterations g=g+1, if reaching maximum iteration Gmax, then population P is exportedgIn optimal solution, otherwise Go to step 3).
CN201410481840.4A 2014-09-19 2014-09-19 A kind of intelligent public transportation dispatching method based on hybrid metaheuristics Active CN104504229B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410481840.4A CN104504229B (en) 2014-09-19 2014-09-19 A kind of intelligent public transportation dispatching method based on hybrid metaheuristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410481840.4A CN104504229B (en) 2014-09-19 2014-09-19 A kind of intelligent public transportation dispatching method based on hybrid metaheuristics

Publications (2)

Publication Number Publication Date
CN104504229A CN104504229A (en) 2015-04-08
CN104504229B true CN104504229B (en) 2018-02-16

Family

ID=52945626

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410481840.4A Active CN104504229B (en) 2014-09-19 2014-09-19 A kind of intelligent public transportation dispatching method based on hybrid metaheuristics

Country Status (1)

Country Link
CN (1) CN104504229B (en)

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105046379A (en) * 2015-09-09 2015-11-11 东南大学 City bus departure timetable optimization method
CN105488581B (en) * 2015-11-13 2019-09-27 清华大学 A kind of transport need amount estimation method based on simulated annealing
CN106448233B (en) * 2016-08-19 2017-12-05 大连理工大学 Public bus network timetable cooperative optimization method based on big data
CN108491968A (en) * 2018-03-17 2018-09-04 北京工业大学 Based on agricultural product quality and safety emergency resources scheduling model computational methods
CN108681824A (en) * 2018-05-24 2018-10-19 湖南智慧畅行交通科技有限公司 Public transit system monthly plan based on heuritic approach is arranged an order according to class and grade algorithm
CN109522584B (en) * 2018-08-31 2022-11-04 南京理工大学 Tramcar departure timetable design method
CN109118023B (en) * 2018-09-21 2022-03-01 北京交通大学 Public transport network optimization method
CN109447557A (en) * 2018-11-05 2019-03-08 安吉汽车物流股份有限公司 Logistic Scheduling method and device, computer readable storage medium
CN110245779A (en) * 2019-05-10 2019-09-17 杭州电子科技大学 A kind of public transport dynamic based on genetic algorithm is dispatched a car method for optimizing scheduling
CN110851247B (en) * 2019-10-12 2023-07-25 华东师范大学 Cost optimization scheduling method for cloud workflow with constraint
CN111161560B (en) * 2019-12-27 2021-05-11 青岛海信网络科技股份有限公司 Bus corridor operation order management method and device
JP7355697B2 (en) * 2020-04-02 2023-10-03 トヨタ自動車株式会社 Vehicle operation control device, operation control method, and transportation system
CN111667114A (en) * 2020-06-05 2020-09-15 鲁东大学 Intelligent route planning recommendation method based on space-time big data fusion
CN111580271B (en) * 2020-06-24 2022-02-18 哈工大机器人(中山)无人装备与人工智能研究院 Self-adaptive aberration correction method and light sheet microscopic imaging device based on same
CN112907935A (en) * 2021-01-31 2021-06-04 安徽达尔智能控制***股份有限公司 Intelligent bus dispatching system and method based on genetic algorithm
CN113128768A (en) * 2021-04-22 2021-07-16 中国电建集团中南勘测设计研究院有限公司 Water-wind-fire short-term optimization scheduling method considering wind-electricity uncertainty
CN113936496B (en) * 2021-12-17 2022-04-01 广东机电职业技术学院 Passenger identification-based intelligent interactive traffic scheduling method and system
CN114358661A (en) * 2022-03-10 2022-04-15 电子科技大学长三角研究院(湖州) Shared bus scheduling method based on pyramid variant group evolution

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101038700A (en) * 2007-04-20 2007-09-19 东南大学 Mixed controlling method of single dot signal controlling crossing
CN102722613A (en) * 2012-05-31 2012-10-10 北京航空航天大学 Method for optimizing electronic component parameters in antenna broadband matching network by adopting genetic-simulated annealing combination
CN103578267A (en) * 2012-07-18 2014-02-12 北京掌城科技有限公司 Bus arrival predication method and system based on bus GPS data
CN103646542A (en) * 2013-12-24 2014-03-19 北京四通智能交通***集成有限公司 Forecasting method and device for traffic impact ranges
CN103678917A (en) * 2013-12-13 2014-03-26 杭州易和网络有限公司 Bus real-time arrival time predicting method based on simulated annealing algorithm

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101038700A (en) * 2007-04-20 2007-09-19 东南大学 Mixed controlling method of single dot signal controlling crossing
CN102722613A (en) * 2012-05-31 2012-10-10 北京航空航天大学 Method for optimizing electronic component parameters in antenna broadband matching network by adopting genetic-simulated annealing combination
CN103578267A (en) * 2012-07-18 2014-02-12 北京掌城科技有限公司 Bus arrival predication method and system based on bus GPS data
CN103678917A (en) * 2013-12-13 2014-03-26 杭州易和网络有限公司 Bus real-time arrival time predicting method based on simulated annealing algorithm
CN103646542A (en) * 2013-12-24 2014-03-19 北京四通智能交通***集成有限公司 Forecasting method and device for traffic impact ranges

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
"Combined Simulated Annealing and Genetic Algorithm Approach to Bus Network Design";Li Liu etal;《Transport systems telematics-10th conference》;20101023;第335-346页 *
"基于改进遗传—模拟退火算法的公交排班优化研究";王庆荣 等;《计算机应用研究》;20120715;第29卷(第7期);第2461-2462页 *
"基于模拟退火遗传算法的多车场车辆调度问题的研究与应用";杨元峰;《中国优秀硕士论文全文数据库 信息科技辑》;20061215(第12期);第I140-62页 *
"基于聚类分析法的城市路网节点交通流量预测研究";赵丽 等;《河北工业大学学报》;20050830;第34卷(第4期);第100-101页第1-3节 *
"改进的遗传—模拟退火算法在公交排班中的应用";黄宏用;《中国优秀硕士论文全文数据库 信息科技辑》;20110915(第9期);第10页第2.4节，第15-16页第3.1-3.2.1节，第17页第3.2.3节，第26-29页第4.1-4.4节，第30-38页第5.2.1-5.3.2节 *

Also Published As

Publication number Publication date
CN104504229A (en) 2015-04-08

Similar Documents

Publication Publication Date Title
CN104504229B (en) A kind of intelligent public transportation dispatching method based on hybrid metaheuristics
CN111178619B (en) Multi-objective optimization method considering distributed power supply and charging station joint planning
CN110245779A (en) A kind of public transport dynamic based on genetic algorithm is dispatched a car method for optimizing scheduling
CN105868942B (en) The orderly charging schedule method of electric car
Yang et al. A bi-objective timetable optimization model incorporating energy allocation and passenger assignment in an energy-regenerative metro system
Chuanjiao et al. Scheduling combination and headway optimization of bus rapid transit
CN108764634A (en) A kind of electric automobile charging station dynamic programming method for considering charge requirement and increasing
CN113077086B (en) Bus synchronous transfer schedule design method for subway junction
Wang et al. Time‐Dependent Electric Vehicle Routing Problem with Time Windows and Path Flexibility
CN106781465A (en) A kind of road traffic Forecasting Methodology
CN111680413B (en) Tramcar timing energy-saving operation optimization method and system based on double-layer algorithm
CN111445094B (en) Express vehicle path optimization method and system based on time requirement
CN115222251A (en) Network taxi appointment scheduling method based on hybrid layered reinforcement learning
Wang et al. Cost-oriented optimization of the location and capacity of charging stations for the electric Robotaxi fleet
CN113033928A (en) Design method, device and system of bus shift scheduling model based on deep reinforcement learning
CN113592419B (en) Rail transit speed and time table optimization method considering passenger flow and energy conservation
CN111222705A (en) Nonlinear charging vehicle path optimization method
Zhao et al. Robust Optimization of Mixed-Load School Bus Route Based on Multi-Objective Genetic Algorithm.
CN113420926A (en) Method for compiling multi-traffic mode transportation capacity collaborative optimization scheme between urban hubs
CN115081153A (en) Bus network design optimization method considering vehicle-mounted capacity and station capacity
Zhu et al. Improved harmony search algorithm for bus scheduling optimization
Zhou et al. Computationally efficient dynamic assignment for on-demand ridesharing in congested networks
Xiao et al. A reinforcement-learning-based bus scheduling model
Xiao et al. A novel bus scheduling model based on passenger flow and bus travel time prediction using the improved cuckoo search algorithm
CN111667087A (en) Bus station-jumping operation method considering pollution emission

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant