CN109872033B - Airport position allocation method - Google Patents
Airport position allocation method Download PDFInfo
- Publication number
- CN109872033B CN109872033B CN201811639136.1A CN201811639136A CN109872033B CN 109872033 B CN109872033 B CN 109872033B CN 201811639136 A CN201811639136 A CN 201811639136A CN 109872033 B CN109872033 B CN 109872033B
- Authority
- CN
- China
- Prior art keywords
- flight
- airport
- positions
- chromosome
- inbound
- 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
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an airport position allocation method, which comprises the following steps: establishing an initial Q table corresponding to all combinations of the inbound flight positions and the departure flight positions in each flight according to the flight information in the pre-allocated flight schedule; screening an entrance station and an exit station combination in each flight by adopting an epsilon greedy search method, and using the entrance station and the exit station combination as a genome for representing each flight; after the flights are arranged in sequence, the corresponding genome forms a chromosome, and the fitness of the chromosome is calculated; updating the Q table according to the return value of each flight; and obtaining a machine position distribution scheme through multiple times of chromosome iteration and Q table updating. The method does not depend on the generation of the initial population, has high calculation convergence speed, and can greatly improve the utilization rate of the airport positions by the airport allocation scheme obtained by the method.
Description
Technical Field
The invention belongs to the field of scheduling allocation, and particularly relates to an airport position allocation method.
Background
The problem of machine position allocation is a complete NP (Non-tertiary polymeric) problem, many domestic airports still adopt manual allocation methods, however, with the continuous expansion of airport scale, the machine position allocation scheme grows exponentially, which brings huge workload to workers, and the realization of intellectualization and convenience of machine position allocation is urgent. The problem of machine position allocation can be solved, airport resources can be fully utilized, airport operation benefit and service quality are improved, and various modern numerical solution algorithms such as a traditional genetic algorithm, a simulated annealing algorithm and the like can be adopted.
The traditional genetic algorithm is a feasible scheme for solving the machine position allocation scheme. However, the traditional genetic algorithm has unstable solving result, weakens the constraints of machine positions and machine types, needs self-correction during crossing and variation, and has the problems of low convergence speed, difficult generation of initial population and the like.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide an airport.
In order to achieve the above object, the present invention provides an airport position allocation method, including:
(1) establishing an initial Q table corresponding to each action of each flight according to the flight information in the pre-allocated flight table; the method specifically comprises the following steps:
(1.1) acquiring flight and airport position information;
the information includes: the method comprises the steps of flight number, model, whether the business is official or not, whether the business is VIP or not, whether the international airline is international airline or not, the time of entering a port, the time of leaving a port, airport scheduling business rule information and model constraint information corresponding to each position;
(1.2) determining a combination of an inbound flight position and an outbound flight position in each flight according to the information;
specifically, each flight is represented by an inbound position and an outbound position, and the inbound position and the outbound position are taken as two genes corresponding to the flight, namely two genes adjacent to each other at a parity position in a chromosome represent one flight;
preferably, for each flight, all possible combinations of genes are referred to as action spaces corresponding to the flight;
(1.3) carrying out Q-table initialization on the combinations of the inbound flight positions and the outbound flight positions in each flight;
preferably, the Q-table initialization process is:
according to airport service hard constraint and soft constraint rules, for the combination of the inbound flight position and the departure flight position in each flight, a positive number is given when the constraint condition is met, and a negative number is given when the constraint condition is not met;
(2) selecting an entrance station and an exit station combination in each flight by adopting an epsilon greedy search algorithm;
preferably, the epsilon greedy search algorithm is: randomly setting a threshold value alpha, screening an entrance position and departure position combination corresponding to the maximum Q value in each flight action when epsilon is less than alpha, otherwise, randomly screening any entrance position and departure position combination in each flight;
(3) constructing a chromosome according to the selected combination of the inbound flight station and the outbound flight station of each flight, and calculating the fitness of the chromosome;
preferably, screening and acquiring combinations of inbound positions and departure positions of each flight according to the epsilon greedy search algorithm, and constructing chromosomes twice as large as the number of flights by using an action-gene mapping table, more specifically, the chromosomes are composed of combinations of inbound positions and departure positions screened by N flights;
preferably, for a chromosome, the set of chromosomes constructed due to the difference in genes in the N genomes is referred to as the state space;
taking a target optimization function contained in the chromosome as a fitness function, namely the departure bridge leaning times, the idle time and the dragging times; however, considering the difference of a plurality of target units and influence amplitudes, the invention adopts a normalization method and a weight coefficient method to solve the problem of multi-target optimization, and the calculation formula of the specific fitness value is as follows: f ═ α Br+βFt+γRn
Wherein:
wherein, FtRepresenting idle time, BrRepresenting the number of bridge abutments, RnRepresenting the dragging times, F representing the fitness value, and alpha, beta and gamma are constants;
(4) checking whether the chromosome meets the constraint condition, and taking the chromosome fitness value meeting the constraint condition as a return value of each flight;
specifically, the obtained chromosome is checked whether to meet the condition constraint through a constraint checking function (Check); if the condition is met, the return value R is equal to F, otherwise, the return value R is equal to-0.01;
(5) updating the Q table according to the return value of each flight;
preferably, the Q-table is updated using the SARSA algorithm:
Qi(st)＝(1-α)Qi(st)+α(R+λ(Qi(st+1)))
wherein: alpha is the learning rate, lambda is the discount coefficient, Qi(St) Represents the Q value corresponding to the combination of the inbound flight position and the departure flight position of the ith flight in t iterations, Qi(St+1) Representing the Q value corresponding to the combination of the inbound flight position and the outbound flight position of the ith flight in (t +1) iterations;
(6) and (5) repeating the steps (2) to (5) according to a preset iteration number, and taking the chromosomes meeting the constraint conditions as a machine position allocation scheme.
Through the technical scheme, compared with the prior art, the invention has the following beneficial effects:
(1) according to the invention, each flight is represented by two genes, so that the readability of the chromosome is enhanced; and further, a constraint test function is adopted for the chromosome to test whether the constraint condition is met, so that the redundancy of the chromosome is avoided.
(2) The method does not need to generate an initial population, and screens Q by an epsilon greedy search algorithmijAnd combining the entry machine position and the departure machine position corresponding to the value to further construct the chromosome, so that the dependence of the traditional genetic algorithm on the initial population is avoided, and meanwhile, the traditional chromosome variation and crossing are not adopted, so that the convergence speed is accelerated.
Drawings
Fig. 1 is a flowchart of a method for allocating a flight level according to the present invention:
fig. 2 is a flowchart of reinforcement learning according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Fig. 1 is a flowchart of a machine position allocation method, fig. 2 is a flowchart specifically showing algorithm processing in the machine position allocation method, and as shown in fig. 1 and fig. 2, the machine position allocation method for an airport provided by the present invention includes:
(1) performing Q-table initialization on all combinations of the inbound flight positions and the departure flight positions in each flight according to the flight information in the pre-allocated flight schedule; the method comprises the following specific steps:
(1.1) acquiring flight and airport position information;
the information includes: the method comprises the steps of flight number, model, whether the business is official or not, whether the business is VIP or not, whether the international airline is international airline or not, the time of entering a port, the time of leaving a port, airport scheduling business rule information and model constraint information corresponding to each position;
(1.2) determining a combination of an inbound flight position and an outbound flight position in each flight according to the information;
specifically, each flight is represented by an inbound position and an outbound position, and the inbound position and the outbound position are taken as two genes corresponding to the flight, namely two genes adjacent to each other at a parity position in a chromosome represent one flight;
preferably, for each flight, all possible combinations of genes are referred to as action spaces corresponding to the flight;
(1.3) carrying out Q-table initialization on the combinations of the inbound flight positions and the outbound flight positions in each flight;
preferably, the Q-table initialization process is:
according to airport service hard constraint and soft constraint rules, for all combinations of inbound positions and departure positions contained in each flight, a positive number is given when the constraint condition is met, and a negative number is given when the constraint condition is not met;
(2) selecting an entrance station and an exit station combination in each flight by adopting an epsilon greedy search algorithm;
preferably, the epsilon greedy search algorithm is: randomly setting a threshold value alpha, screening an entrance position and departure position combination corresponding to the maximum Q value in each flight action when epsilon is less than alpha, otherwise, randomly screening any entrance position and departure position combination in each flight;
(3) constructing a chromosome according to the selected combination of the inbound flight station and the outbound flight station of each flight, and calculating the fitness of the chromosome;
preferably, screening and acquiring combinations of inbound positions and departure positions of each flight according to the epsilon greedy search algorithm, and constructing chromosomes with the size being twice of the number of flights by using an action-gene mapping table, more specifically, the chromosomes are composed of combinations of inbound positions and departure positions screened by N flights;
preferably, for a chromosome, the set of chromosomes constructed due to the difference in genes in the N genomes is referred to as the state space;
taking a target optimization function contained in the chromosome as a fitness function, namely the departure bridge leaning times, the idle time and the dragging times; however, considering the difference of a plurality of target units and influence amplitudes, the invention adopts a normalization method and a weight coefficient method to solve the problem of multi-target optimization, and the calculation formula of the specific fitness value is as follows: f ═ α Br+βFt+γRn
Wherein:
wherein, FtRepresenting idle time, BrRepresenting the number of bridge abutments, RnRepresenting the dragging times, F representing the fitness value, and alpha, beta and gamma are constants;
(4) checking whether the chromosome meets the constraint condition, and taking the chromosome fitness value meeting the constraint condition as a return value of each flight;
checking whether the obtained chromosome meets the condition constraint through a constraint checking function (Check); if the condition is met, the return value R is equal to F, otherwise, the return value R is equal to-0.01;
(5) updating the Q table according to the return value of each flight;
preferably, the Q-table is updated using the SARSA algorithm:
Qi(st)＝(1-α)Qi(st)+α(R+λ(Qi(st+1)))
wherein: alpha is the learning rate, lambda is the discount coefficient, Qi(St) Represents the Q value corresponding to the combination of the inbound flight position and the departure flight position of the ith flight in t iterations, Qi(St+1) Representing the Q value corresponding to the combination of the inbound flight position and the outbound flight position of the ith flight in (t +1) iterations;
(6) and (5) repeating the steps (2) to (5), and when the iteration times is larger than the preset iteration times, terminating the algorithm, and taking the chromosomes meeting the constraint conditions as a machine position distribution scheme.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (5)
1. An airport position allocation method, comprising:
(1) performing Q table initialization on all combinations of the inbound flight positions and the departure flight positions in each flight according to flight information in a pre-allocated flight schedule, wherein the combinations of the inbound flight positions and the departure flight positions correspond to Q values one by one, when the combinations of the inbound flight positions and the departure flight positions meet constraint conditions, positive initial values of the corresponding Q values are given, otherwise, negative initial values of the corresponding Q values are given, and the Q values are combined to form the Q table;
(2) screening an inbound flight position and an outbound flight position combination in each flight according to the comparison between a set threshold value and the Q value of each flight;
(3) taking an inbound flight position and an outbound flight position as the genome of each flight, constructing chromosomes corresponding to the genome according to the arrangement sequence set by each flight, and calculating the fitness of the chromosomes;
(4) whether the chromosome meets the constraint condition is checked, the fitness value of the chromosome meeting the constraint condition is used as a return value corresponding to each flight, and the fitness value of the chromosome is as follows: f ═ α Br+βFt+γRnWherein:
Ftrepresenting idle time, BrRepresenting the number of bridge abutments, RnRepresenting the dragging times, F representing the fitness value, and alpha, beta and gamma are constants;
(5) updating the Q table according to the return value of each flight;
(6) and (5) repeating the steps (2) to (5) according to the preset iteration number, and taking the chromosome which finally meets the constraint condition as a machine position allocation scheme.
2. The method of assigning airport stands according to claim 1, wherein step (1) comprises the steps of:
(1.1) acquiring flight and airport position information;
(1.2) determining the combination of the inbound flight position and the outbound flight position of each flight according to the acquired information;
and (1.3) carrying out Q table initialization on the combination of the inbound flight position and the outbound flight position in each flight.
3. The method of airport terminal assignment as claimed in claim 2 wherein said flight and airport terminal information comprises: the method comprises the steps of flight number, model, whether the model is official or not, whether the model is VIP or not, whether the model is international airline or not, the time of entering a port, the time of leaving the port, airport scheduling business rule information and model constraint information corresponding to each position.
4. The method of airport airfield allocation of claim 1 or 3 wherein the Q table initialization procedure is:
and according to the airport service hard constraint and soft constraint conditions, for each flight action space, a positive number is given when the constraint conditions are met, and a negative number is given when the constraint conditions are not met.
5. The method of assigning airport positions of claim 1, wherein the assignment method of each flight return value in step (4) is:
the obtained chromosome is checked whether to accord with the constraint condition through a constraint check function; if the condition is met, the report value R is equal to F, otherwise, the report value R is equal to-0.01.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811639136.1A CN109872033B (en) | 2018-12-29 | 2018-12-29 | Airport position allocation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811639136.1A CN109872033B (en) | 2018-12-29 | 2018-12-29 | Airport position allocation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109872033A CN109872033A (en) | 2019-06-11 |
| CN109872033B true CN109872033B (en) | 2021-02-26 |
Family
ID=66917319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811639136.1A Active CN109872033B (en) | 2018-12-29 | 2018-12-29 | Airport position allocation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109872033B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110264005B (en) * | 2019-06-21 | 2022-02-15 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN110443448B (en) * | 2019-07-01 | 2022-03-29 | 华中科技大学 | Bidirectional LSTM-based airplane position classification prediction method and system |
| CN110549896B (en) * | 2019-08-28 | 2022-08-02 | 哈尔滨工程大学 | Charging station selection method based on reinforcement learning |
| CN111079274B (en) * | 2019-12-04 | 2024-04-09 | 深圳市机场股份有限公司 | Intelligent allocation method for machine position, computer device and storage medium |
| CN111178004B (en) * | 2019-12-30 | 2023-12-26 | 北京富通东方科技有限公司 | Regular digital representation method for airport stand resource allocation |
| CN111951145B (en) * | 2020-08-12 | 2022-06-10 | 青岛民航凯亚系统集成有限公司 | GA-DQN-based shutdown position distribution method |
| CN117237241B (en) * | 2023-11-15 | 2024-02-06 | 湖南自兴智慧医疗科技有限公司 | Chromosome enhancement parameter adjustment method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1153337A (en) * | 1997-07-30 | 1999-02-26 | Fujitsu Ltd | Optimization method and device using object-oriented genetic algorithm |
| CN107679669A (en) * | 2017-10-19 | 2018-02-09 | 云南大学 | A kind of airport aircraft gate dispatching method and system based on meta-heuristic method |
-
2018
- 2018-12-29 CN CN201811639136.1A patent/CN109872033B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1153337A (en) * | 1997-07-30 | 1999-02-26 | Fujitsu Ltd | Optimization method and device using object-oriented genetic algorithm |
| CN107679669A (en) * | 2017-10-19 | 2018-02-09 | 云南大学 | A kind of airport aircraft gate dispatching method and system based on meta-heuristic method |
Non-Patent Citations (3)
| Title |
|---|
| RLGA: 一种基于强化学习机制的遗传算法;王本年等;《电子学报》;20060531;第856-860页 * |
| 基于遗传算法的机场机位分配策略;田晨等;《计算机工程》;20050228;第186-188页 * |
| 民航机场机位分配问题的遗传算法研究;沈洋;《中国管理科学》;20061031;第105-108页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109872033A (en) | 2019-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109872033B (en) | Airport position allocation method | |
| CN107453919B (en) | Complex network node importance evaluation method and system | |
| CN107169628B (en) | Power distribution network reliability assessment method based on big data mutual information attribute reduction | |
| Gan et al. | Joint optimization of maintenance, buffer, and spare parts for a production system | |
| CN112070355A (en) | Distribution scheduling method for airport ferry vehicle | |
| CN113536081B (en) | Data center data management method and system based on artificial intelligence | |
| CN105929690A (en) | Flexible workshop robustness scheduling method based on decomposition multi-target evolution algorithm | |
| CN109768540B (en) | Power distribution network power failure optimization scheduling method based on big data analysis | |
| CN109710406A (en) | Data distribution and its model training method, device and computing cluster | |
| CN104899101B (en) | Software testing resource dynamic allocation method based on multi-target difference evolution algorithm | |
| CN111950910B (en) | Airport guarantee vehicle task scheduling method based on DBSCAN-GA | |
| CN108108883B (en) | Clustering algorithm-based vehicle scheduling network elastic simplification method | |
| CN110019625B (en) | Text standard address spatialization method and device and computer readable storage medium | |
| CN113887748B (en) | Online federal learning task allocation method and device, and federal learning method and system | |
| US10313457B2 (en) | Collaborative filtering in directed graph | |
| KR101565694B1 (en) | Artificial bee colony clustering design method and system in wireless sensor network | |
| Sun et al. | Wireless sensor network node localization based on genetic algorithm | |
| CN117236640A (en) | Method, device, terminal and medium for optimizing equipment configuration scheme of low-voltage distribution network | |
| CN115907079B (en) | Airspace traffic flow prediction method based on attention space-time diagram convolutional network | |
| CN109377111A (en) | Job scheduling method and device based on modified-immune algorithm | |
| Chiu et al. | Novel hybrid approach with elite group optimal computing budget allocation for the stochastic multimodal problem | |
| CN111917589B (en) | Power communication network resource backup method and related device | |
| CN112948087A (en) | Task scheduling method and system based on topological sorting | |
| CN114265630A (en) | Resource sharing perception online task unloading method | |
| CN113761666B (en) | Automatic processing method for aircraft quality characteristic data |
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 |