CN113065692A - Prediction method for service time of airport water truck - Google Patents
Prediction method for service time of airport water truck Download PDFInfo
- Publication number
- CN113065692A CN113065692A CN202110300643.8A CN202110300643A CN113065692A CN 113065692 A CN113065692 A CN 113065692A CN 202110300643 A CN202110300643 A CN 202110300643A CN 113065692 A CN113065692 A CN 113065692A
- Authority
- CN
- China
- Prior art keywords
- time
- distribution
- service time
- airport
- grouping
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 104
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000009826 distribution Methods 0.000 claims abstract description 150
- 239000013505 freshwater Substances 0.000 claims abstract description 63
- 238000004364 calculation method Methods 0.000 claims abstract description 12
- 238000007689 inspection Methods 0.000 claims abstract description 7
- 238000000546 chi-square test Methods 0.000 claims description 17
- 238000012216 screening Methods 0.000 claims description 11
- 238000004140 cleaning Methods 0.000 claims description 8
- 230000010006 flight Effects 0.000 claims description 5
- 238000007476 Maximum Likelihood Methods 0.000 claims description 4
- 238000005315 distribution function Methods 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 2
- 238000011160 research Methods 0.000 abstract description 11
- 238000011156 evaluation Methods 0.000 abstract description 3
- 230000009286 beneficial effect Effects 0.000 abstract 1
- 230000000717 retained effect Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000009795 derivation Methods 0.000 description 1
- 239000003651 drinking water Substances 0.000 description 1
- 235000020188 drinking water Nutrition 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Probability & Statistics with Applications (AREA)
- Game Theory and Decision Science (AREA)
- Algebra (AREA)
- Development Economics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for predicting the service time of an airport waterwheel, which comprises the following steps: the method comprises four steps of preliminary statistics of the service time of the fresh water vehicles, grouping of the service time of the fresh water vehicles, probability distribution fitting and inspection of the service time of the fresh water vehicles and prediction of the service time of the fresh water vehicles. The invention provides a grouping method of clear water vehicle operation guarantee data, aiming at various types of clear water vehicle operation guarantee data, the clear water vehicle operation guarantee data can be grouped indiscriminately, and follow-up calculation and prediction can be guaranteed within feasible time. Based on a grouping method, a classification, initial prediction and final prediction fresh water vehicle service time prediction method is further provided, and compared with the traditional experience probability prediction method, the prediction accuracy of the fresh water vehicle service time is greatly improved. The method has higher prediction accuracy, is beneficial to airport guarantee capability evaluation and airport guarantee vehicle scheduling research based on the service time of the clear water vehicle, ensures that the research content has stronger practical significance, and meets the requirement of practical engineering application.
Description
Technical Field
The invention relates to a method for predicting the service time of a clear water vehicle in an airport, belonging to the technical field of prediction of the service time of the clear water vehicle.
Background
The airport clean water vehicle is one kind of airport guarantee vehicle, provides the service of drinking water filling for the aircraft, plays crucial effect in the guarantee process of aviation shift. At present, research relating to guaranteeing vehicle service time is mainly applied to two major fields: airport security capability evaluation and airport security vehicle scheduling research. In conventional research, researchers have used almost all a fixed value or an average value of guaranteed vehicle service times as the guaranteed vehicle service time. Few studies have used a triangular or gamma distribution to represent guaranteed vehicle service time, but have not verified the scientificity and significance of the distribution.
For the collection of the service time data of the fresh water vehicle, only a few airports are available in China, such as Beijing Daxing International airport, Beijing capital International airport, Shanghai Pudong International airport and Guangzhou white cloud International airport. Meanwhile, the current research on the service time of the fresh water truck is basically blank. The current research about airport clean water vehicle service time mainly has the following defects: (1) the fixed value or the average value is directly used as the service time of the clear water vehicle, so that the clear water vehicle is inconsistent with objective facts; (2) the distribution is directly used as the distribution of the service time of the water truck, but the reasonability and the scientificity are lacked only by a manual direct observation method and simple probability distribution fitting; (3) due to the defects of the previous research, the service time of the clear water vehicle cannot be reasonably predicted. If the service time of all the water trucks is represented by using a certain probability distribution as in the previous research, and then the prediction is carried out on a certain confidence level, the prediction accuracy is low, and the requirements of actual engineering application cannot be met.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method for predicting the service time of the fresh water vehicle in the airport can be suitable for different types of fresh water vehicle guarantee data forms, and compared with the original related research, the method for predicting the service time of the fresh water vehicle in the airport improves the prediction accuracy rate of the service time of the fresh water vehicle. The method has higher prediction accuracy, is favorable for airport guarantee capability evaluation and airport guarantee vehicle scheduling research based on the service time of the clear water vehicle, has stronger practical significance, and meets the requirement of practical engineering application.
The invention adopts the following technical scheme for solving the technical problems:
a prediction method for airport fresh water vehicle service time comprises the following steps:
and 8, obtaining a final prediction result according to the initial prediction fresh water truck service time lower limit and the initial prediction fresh water truck service time upper limit which correspond to all the alternative groups respectively.
As a preferred scheme of the present invention, step 1, the origin airport, i.e. the airport to be predicted; the planned departure time is the planned departure time of a departure place airport wheel-withdrawing gear which is declared in advance by departure flights, and the time precision is 5 minutes; the actual departure time is the actual time for departure flights to withdraw the wheel gear from the origin airport in the actual operation, and the time precision is 1 minute; the service starting time and the service ending time of the clear water vehicle are respectively the time points when the departure flight is actually running, the airport clear water vehicle is actually the time point when the departure flight starts to fill clear water and ends to fill clear water, and the time precision is 1 minute.
As a preferred scheme of the present invention, the step 1 of cleaning and calculating the guarantee data specifically includes:
for each piece of guarantee data, if at least one of the fresh water truck starting service time and the fresh water truck ending service time of the piece of data is in an error form, removing the piece of data; and (3) subtracting the service starting time of the clean water vehicle from the service ending time of the clean water vehicle of the data to obtain the service time of the clean water vehicle of the data, wherein the time precision of the service time of the clean water vehicle is 1 minute.
As a preferred scheme of the present invention, the number of grouping schemes for guaranteeing data division under each combination in step 3, i.e. the complexity, is calculated by the following formula:
wherein D represents the complexity of the guaranteed data partitioning calculated when m attributes are selected from the remaining 7 attributes and combined, m represents the total number of all attributes used for calculating the complexity, n is 1, …, m, anIndicates the total number of data with the n-th attribute removed, i is 1, …, an；
Attributes requiring deduplication are planned departure time and driving time, converting planned departure time with a time precision of 5 minutes into planned departure time with a precision of 1, 2, 3, 4, 6, 8, 12, or 24 hours, and converting driving time with a time precision of 1 minute into driving time with a precision of 1, 2, 3, 4, 6, 8, 12, or 24 hours.
As a preferred embodiment of the present invention, in step 5, probability distribution fitting and probability distribution inspection are performed on each grouping scheme to obtain a final probability distribution of the grouping scheme, which specifically includes:
aiming at a grouping scheme, 20 kinds of probability distribution are adopted to respectively fit the service time of the water truck, and a maximum likelihood method is used for solving each probability distribution parameter, wherein the 20 kinds of probability distribution comprise 4 types: single-parameter probability distribution, two-parameter probability distribution, three-parameter probability distribution and four-parameter probability distribution; the single parameter probability distribution includes: exponential distribution, poisson distribution, and rayleigh distribution; the two-parameter distribution includes: bernbaum sanded distribution, extremum distribution, gamma distribution, semi-normal distribution, inverse gaussian distribution, logical distribution, log-normal distribution, mid-upper distribution, normal distribution, leis distribution, and weibull distribution; the three-parameter distribution includes: a Bohr distribution, a generalized extreme distribution, a generalized pareto distribution, and a t-position scale distribution; the four parameter distribution includes: stable distribution; and (4) carrying out chi-square test on the 20 probability distribution fitting results of the grouping scheme, and selecting the probability distribution with the highest goodness of fit as the final probability distribution of the grouping scheme.
As a preferred embodiment of the present invention, the calculation formula of the initial prediction in step 7 is:
wherein d represents a predicted correct probability value under a preset prediction precision, b represents a predicted lower limit of the service time of the fresh water truck, Δ c represents a preset precision of the predicted service time of the fresh water truck, b + Δ c represents a predicted upper limit of the service time of the fresh water truck, and p (x) represents a probability distribution function of the alternative grouping;
when d isbWhen the value is constantly larger than 0, the value of b is the maximum value of the definition domain minus delta c; when d isbWhen the value is constantly less than 0, the value of b is the minimum value of the definition domain; when d isb' when it is not more than and not less than 0, b is dbMaximum value of d when' equals 0.
As a preferred embodiment of the present invention, the calculation formula of step 8 is:
wherein, F0And F1Respectively representing the lower and upper limits of the final prediction result,
and
respectively representing the lower and upper limits of the initial prediction of the jth candidate packet, GjRespectively representing the goodness of fit of the probability distribution corresponding to the jth candidate group, and k representing the number of all the candidate groups.
Compared with the prior art, the invention adopting the technical scheme has the following technical effects:
1. the invention provides a grouping method of clear water vehicle operation guarantee data, aiming at various types of clear water vehicle operation guarantee data, the clear water vehicle operation guarantee data can be grouped indiscriminately, and follow-up calculation and prediction can be guaranteed within feasible time.
2. The invention further provides a clear water vehicle service time prediction method of classification, initial prediction and final prediction based on a grouping method, and compared with the traditional empirical probability prediction method, the prediction accuracy of the clear water vehicle service time is greatly improved.
Drawings
FIG. 1 is a flow chart of a method for forecasting service time of an airport waterwheel of the present invention.
FIG. 2 is a partial probability distribution fit of a group of the service times for a clearwater vehicle of the present invention.
FIG. 3 is a probability distribution goodness of fit for a group of service times for a clearwater vehicle of the present invention.
FIG. 4 is a comparison of the accuracy of the present invention prediction method versus the conventional prediction method for predicting the service time of a fresh water truck.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
Fig. 1 is a flow chart of a method for predicting service time of an airport waterwheel according to the present invention. The method comprises the following specific steps:
The preliminary statistics of the service time of the water truck are derived from the operation guarantee data of the water truck in the airport. Common airport clean water vehicle operation support data generally include: departure flight number, model class, origin airport, destination airport, planned departure time, actual departure time, clean water vehicle start service time, and clean water vehicle end service time.
The departure flight number of the operation guarantee data of the airport clean water vehicle is the call number of the departure flight corresponding to the data. The "model" of the guarantee data is a specific model code of the departing flight, including but not limited to: a3XX (air passenger series), B7XX (boeing series), and the like. The 'model class' of the guarantee data is a wake flow type to which the departure flight model belongs, and comprises six classes of A/B/C/D/E/F. The "origin airports" of the support data are all identical, indicating that all departure flights depart from the airport under study, and are represented by the ICAO (International Civil Aviation organization) four-character code for the airport under study. The "destination airport" of the assurance data is the ICAO four-word code of the other airports at which the airport under study navigates. The "planned departure time" of the guarantee data is the planned time for removing the gear from the origin airport declared in advance by the departure flight, and is expressed as an integral multiple of 5 minutes (i.e., 5 minutes in time accuracy). The "actual departure time" of the guarantee data is the actual time for the departure flight to withdraw the gear from the origin airport in the actual operation, and is expressed as an integral multiple of 1 minute (i.e. time precision of 1 minute). The 'clear water vehicle starting service time' and 'clear water vehicle ending service time' of the guarantee data are time points when the departure flight is actually running and the airport clear water vehicle actually starts to fill clear water for the departure flight and ends the operation of filling clear water, and are expressed as integral multiples of 1 minute (namely time precision is 1 minute).
The preliminary statistical process of the service time of the fresh water truck comprises two substeps of data cleaning and data calculation. Data cleaning means that for each piece of guarantee data, if the 'fresh water truck starting service time' or 'fresh water truck ending service time' of the data is in an error form (including blank and non-time data forms), the data is removed. And (3) data calculation, namely subtracting the service starting time of the clear water vehicle from the service ending time of the clear water vehicle of each piece of guarantee data subjected to data cleaning to obtain the service starting time of the clear water vehicle of the data. Since the time accuracy of the "fresh water truck service start time" and the "fresh water truck service end time" is 1 minute, the time accuracy of the "fresh water truck service time" is also 1 minute.
On the basis of the step 1, the grouping process of the service time of the water truck comprises three substeps of determining initial grouping attributes, screening complexity and determining grouping.
2.1 determining initial grouping attributes the purpose of determining preliminary attributes for the grouped assurance data is to preliminarily determine the individual or multiple attributes for the grouped assurance data. As described in step 1, the safeguard data generally has 9 attributes, and 7 attributes other than "actual departure time" and "end service time of a clean water truck" may be determined as the initial grouping attributes. The reason why the "actual departure time" and the "clear water truck end service time" are excluded is that: the purpose of the invention is to predict the service time of the fresh water truck, and the time nodes of the 'actual departure time' and the 'end service time' of the fresh water truck are all behind the 'start service time of the fresh water truck', so the attributes cannot be used for prediction analysis. By extension, if an attribute of the provisioning data (whether or not of the time type) is determined later than the "fresh water truck start service time", that attribute cannot be used to determine the initial packet attribute.
2.2 complexity screening aims at calculating the number of schemes for guaranteeing data division according to the determined single or multiple combinations of the initial grouping attributes, and screening grouping modes with reasonable number of schemes. If the number of schemes is too large (sometimes the number of numerical digits is as high as tens of digits, hundreds of digits or even higher), the computer cannot complete the grouping operation within a feasible time, and is less likely to perform the subsequent steps within the feasible time. The complexity calculation is shown in equation (1): d represents the complexity, m represents the total number of all attributes used to compute the complexity, n represents a sequence of certain attributes, anRepresents the total number of data after the n-th attribute is removed, and i represents a certain data sequence after the n-th attribute is removed. If the complexity is larger than a preset complexity threshold value, screening the scheme; otherwise, the method is reserved. For attributes in time or numeric form, its accuracy can be adjusted (e.g., a "planned departure time" with an accuracy of 5 minutes is converted to hours of a "planned departure time" with an accuracy of 1/2/3/4/6/8/12/24 hours, such that 288 deduplicated data become 24/12/8/6/4/3 @duringthe day2/1 deduplication data); for other forms of attributes, it remains unchanged.
2.3 determining groupings is an operation based on determining initial grouping attributes and complexity screening, with the goal of ensuring that the distribution of fresh water truck service times between mutually independent groupings differs significantly at some confidence level. The invention uses the traditional chi-square test method to carry out chi-square test of twice crossing (two groups are theoretical values or observed values of each other) on the empirical probability distribution of the groups between every two independent groups, and is used for testing the necessity of the groups. If the guaranteed data quantity of one group in two independent groups is less than 40, the group is directly removed, because the packet necessity cannot be judged by too small data quantity, and the prediction of the service time of the water truck cannot be carried out on the basis of the packet necessity. Both independent packets are retained if their two-fold cross chi-square test both show significant non-identity to the same empirical probability distribution at a 99.9% confidence level; instead, the two independent packets are combined into one packet.
And (3) performing probability distribution fitting and inspection on the grouping result of the step 2 on the basis of the step 1 and the step 2 respectively. The probability distribution fitting and checking of the service time of the water truck comprises the probability distribution fitting and the probability distribution checking, and comprises two substeps.
3.1 probability distribution fitting
In general, the precision of the service time of the water truck in step 1 is 1 minute, so when the data amount is sufficient, the frequency of the service time of the water truck should be counted by taking 1 minute as an interval. If the data volume is small, the service time frequency of the clean water vehicles at most intervals is 0, and the intervals should be properly enlarged.
The method for solving the parameters of the probability distribution uses a maximum likelihood method by using 20 kinds of probability distribution common in engineering to respectively fit the service time of the fresh water truck. The 20 distributions common in engineering include 4 major classes: single parameter probability distribution, two parameter probability distribution, three parameter probability distribution, and four parameter probability distribution. The single parameter probability distribution includes: the total number of the exponential distribution, the Poisson distribution and the Rayleigh distribution is 3; the two-parameter distribution includes: a total of 12 kinds of Bernbaum Mordes distribution, extreme value distribution, gamma distribution, semi-normal distribution, inverse Gaussian distribution, logic distribution, logarithmic normal distribution, middle-upper distribution, normal distribution, Lais distribution and Weibull distribution; the three-parameter distribution includes: 4 kinds of distribution in total are Bohr distribution, generalized extreme value distribution, generalized pareto distribution and t position scale distribution; the four parameter distribution includes: the distribution is stable, and the total number is 1.
3.2 probability distribution test
Then, chi-square test is carried out on various probability distribution fitting results of each group, and the probability distribution with the highest goodness of fit is selected as the final probability distribution of the group.
And (3) predicting the service time of the clear water vehicle of certain guarantee data on the basis of the steps 1 to 3. The prediction of the service time of the fresh water truck comprises three substeps of classification, initial prediction and final prediction.
4.1 the classification is according to the departure flight number, model class, origin airport, destination airport, planned departure time and clear water vehicle starting service time of the guarantee data to be predicted, if the attribute of a certain group in the step 3 is in accordance with all the attributes of the guarantee data, the guarantee data is classified into the group, and the group is taken as one of the alternative groups of the guarantee data.
4.2 the initial prediction is to perform initial prediction calculation on the service time of the fresh water vehicles of the guarantee data according to preset prediction precision for each alternative group of the guarantee data. The initial prediction calculation for a candidate packet is shown in equation (2): d represents a predicted correct probability value under a preset prediction precision, b represents a predicted lower limit of the service time of the fresh water truck, Δ c represents a preset precision of the predicted service time of the fresh water truck, and b + Δ c represents a predicted fresh water truckThe waterwheel service time upper bound, p (x), represents the probability distribution function for this alternative grouping. In order to improve the accuracy of prediction, the maximum value of d needs to be obtained, and the derivation process is shown in formula (3): when d isbWhen the value is constantly larger than 0, the value of b is the maximum value of the definition domain minus delta c; when d isbWhen the value is constantly less than 0, the value of b is the minimum value of the definition domain; when d isb' when it is not more than and not less than 0, b is dbMaximum value of d when' equals 0.
4.3 the final prediction is an initial prediction of all the alternative groupings based on the assurance data, which is expected to yield a final prediction result. According to the preset precision Δ c, the upper and lower limits of the final prediction are solved as shown in formula (4): f0And F1Respectively representing the lower and upper limits of the final prediction,
and
respectively representing the lower and upper limits, G, of the initial prediction of the jth candidate packet of the safeguard datajChi-squared test significant correlation probabilities (considered as probability distribution goodness of fit) respectively representing initial predicted probability distributions of jth candidate packets of the safeguard data, k represents the number of candidate packets of the safeguard data. Due to the fact that for any
And
are all provided with
So F1-F0Δ c, i.e., the accuracy of the final prediction and the accuracy of the initial prediction are both Δ c.
With reference to fig. 2 to 4, the following embodiments are described in detail by taking part of the fresh water truck support data of the great international airport in beijing as an example, and the method for predicting the service time of the fresh water truck mainly includes the following steps:
Taking the example of the international airport in great Beijing from 2019 to 2020 5 months, the operation guarantee data of the common airport clean water vehicle are shown in Table 1 and comprise: departure flight number, model class, origin airport, destination airport, planned departure time, actual departure time, clean water vehicle start service time, and clean water vehicle end service time.
Table 1 data format sample
The departure flight number of the piece of guarantee data is CA757, the machine type is B737, and the machine type is class C; the origin airport is the Beijing great-rise International airport (ZBAD) and the destination airport is the Thailand Gallman International airport (VTBD). The "planned departure time" of this piece of guarantee data is the planned departure time of the departure flight CA757 at the origin airport, and the value is 7 hours and 50 minutes (time accuracy 5 minutes). The "actual departure time" of this piece of guarantee data is the actual gear-off time of the departure flight CA757 at the origin airport in actual operation, and has a value of 7 hours and 51 minutes (time accuracy 1 minute). The 'clear water vehicle starting service time' and 'clear water vehicle ending service time' of the piece of guarantee data are the time points when the departure flight CA757 is actually running and the airport clear water vehicle actually starts to fill clear water for the departure flight and ends to fill clear water, and the numerical values are 22 minutes (time precision 1 minute) and 27 minutes (time precision 1 minute) at 6 hours respectively.
1) Data cleansing
Before data cleansing, the number of guaranteed data pieces for the example is 31076. For each piece of guarantee data, if the "fresh water truck start service time" or "fresh water truck end service time" of the data is in an error form (including blank and non-time data forms), the piece of data is removed. After data washing, the number of the data pieces of the example is 15933.
2) Data computation
And (3) data calculation, namely subtracting the service starting time of the clear water vehicle from the service ending time of the clear water vehicle of each piece of guarantee data subjected to data cleaning to obtain the service starting time of the clear water vehicle of the data. Taking table 1 as an example, the service time of the clean water vehicle of the piece of guarantee data is 5 minutes. Since the time accuracy of the "fresh water truck service start time" and the "fresh water truck service end time" is 1 minute, the time accuracy of the "fresh water truck service time" is also 1 minute.
1) Determining initial grouping
The purpose of determining the initial grouping attributes is to preliminarily determine a single or multiple attributes of the provisioning data for the grouping. As described in step 1, the safeguard data generally has 9 attributes, and 7 attributes other than "actual departure time" and "end service time of a clean water truck" may be determined as the initial grouping attributes.
2) Complexity screening
The purpose of complexity screening is to calculate the number of schemes for guaranteeing data division according to the determined single or multiple combinations of the initial grouping attributes, and screen grouping modes with reasonable number of schemes.
Taking a single attribute of a model as an example, the number of effective guarantee data pieces of the output example in the step 1 is 15933, and different models have 14 types. From equation (1), there are 65535 total grouping schemes, i.e., 65535 complexity. If the predetermined complexity threshold is 1000, the complexity of the attribute exceeds the threshold, so it is filtered out.
Taking a single attribute of 'model large class' as an example, the number of effective guarantee data pieces output in the step 1 is 15933, and 5 types of different model large classes exist. From equation (1), a total of 31 grouping schemes can be derived, i.e., a complexity of 31. If the preset complexity threshold is 1000, the complexity of the attribute does not exceed the threshold, and is retained.
Taking the multiple attributes of "model classes" and "planned departure times (24 time slices in total, i.e., the precision is 1 hour, i.e., the time slices are divided into 0 hour to 1 hour, 1 hour to 2 hours, etc.)" as an example, there are 5 types of different model classes and 24 types of different planned departure times. From equation (1), there are 520093665 grouping schemes, i.e. 520093665 complexity, which can be derived from 31 × 16777215. If the predetermined complexity threshold is 1000, the complexity of the attribute exceeds the threshold, so it is filtered out.
Taking the multiple attributes of "model classes" and "planned departure times" (accuracy 6 hours, i.e., divided into 0 hour to 6 hours, 6 hours to 12 hours, 12 hours to 18 hours, and 18 hours to 24 hours) as an example, there are 5 types of different model classes and 4 types of different planned departure times. From equation (1), a total of 31 × 15-465 grouping schemes can be obtained, i.e., complexity 465. If the preset complexity threshold is 1000, the complexity of the attribute does not exceed the threshold, and is retained.
3) Determining a packet
Determining the groupings is based on determining initial grouping attributes and a complexity screening operation in order to ensure that the distribution of the fresh water truck service times between the independent groupings differs significantly at some confidence level.
Taking the multi-attribute grouping of the guarantee data according to the 'model class' and the 'planned departure time' as an example, 465 grouping schemes are counted. First, 30 packets out of 465 packets have a guaranteed data amount less than 40, and are directly removed, leaving 435 packets. Using the chi-squared test method, two independent groups are retained if chi-squared tests of two crossings of the two independent groups (the two groups being either theoretical or observed values of each other) both show significant non-identity to the same empirical probability distribution at a 99.9% confidence level; instead, the two independent packets are combined into one packet. Through screening by the chi-square test method, 435 packets are combined into 36 packets.
And (3) performing probability distribution fitting and inspection on the grouping result of the step 2 on the basis of the step 1 and the step 2 respectively. And (3) for the grouping result in the step (2), fitting the service time data of the water wagons by using 20 kinds of probability distribution common in engineering, respectively fitting the service time of the water wagons in each group, and solving each probability distribution parameter by using a maximum likelihood method. Partial probability distribution fit for a certain set is shown in fig. 2, and significance of chi-squared test is ranked from high to low as shown in fig. 3. The higher significance of a certain probability distribution of fig. 3 indicates that the better the fit of such probability distribution. In the packet analysis of fig. 2 and 3, the 16 th probability distribution (boolean distribution) with the highest goodness of fit is selected as the final probability distribution of the packet.
1) Classification
Taking the guarantee data of table 1 as an example, if there is some grouping manner: for example, grouped in multiple attributes of "model class" and "planned departure time (accuracy 6 hours, i.e., divided into 0 hour to 6 hours, 6 hours to 12 hours, 12 hours to 18 hours, and 18 hours to 24 hours)", the safeguard data should be classified into a group of "model class" of C and "planned departure time" of 6 hours to 12 hours and other groups (if the attributes of a certain group conform to all the attributes of the safeguard data).
2) Initial prediction
According to the above classification results, the safeguard data of table 1 has a plurality of alternative groupings, and according to formula (3), a plurality of initial predicted lower and upper limits can be output according to different preset precisions Δ c.
3) Final prediction
Based on the above-mentioned multiple lower and upper limits of the initial prediction, the final prediction results at different preset accuracies Δ c can be calculated according to equation (4).
According to final predicted F0And F1(i.e., lower and upper limits) if the warranty dataThe service time of the clean water vehicle is more than or equal to F0And is less than or equal to F1If the service time of the fresh water vehicle is predicted correctly, the guarantee data is related to the preset precision delta c; if the service time of the fresh water truck of the guarantee data is less than F0Or greater than F1The guarantee data is wrong with respect to the prediction of the service time of the fresh water vehicle with the preset precision deltac. According to the times of correct prediction and wrong prediction, the accuracy of the output prediction is high and low.
The clear water vehicle guarantee data of Beijing great-rise international airport 2020 and 6 months is used as prediction test data, the traditional clear water vehicle service time prediction method based on experience probability distribution is firstly used for prediction, then the prediction method is used for prediction, and the comparison of prediction accuracy rates is shown in a table 2 and a figure 4.
TABLE 2 comparison of prediction accuracy of the present invention versus conventional prediction methods
The above embodiments are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modifications made on the basis of the technical scheme according to the technical idea of the present invention fall within the protection scope of the present invention.
Claims (7)
1. A prediction method for the service time of an airport waterwheel is characterized by comprising the following steps:
step 1, acquiring operation guarantee data of the clear water vehicle of the airport aiming at the airport to be predicted, wherein the operation guarantee data comprises 9 attributes including departure flight number, machine type category, origin airport, destination airport, planned departure time, actual departure time, clear water vehicle starting service time and clear water vehicle ending service time, and cleaning and calculating the guarantee data;
step 2, regarding 9 attributes of the guarantee data, wherein the actual departure time and the determined time of the service ending time of the clean water truck are both later than the service starting time of the clean water truck, and the remaining 7 attributes after the two attributes are removed are used as initial grouping attributes;
step 3, combining the remaining 7 attributes, calculating the number of grouping schemes for guaranteeing data division under each combination, namely complexity, and setting a complexity threshold, if the number of the grouping schemes calculated by a certain combination is smaller than the complexity threshold, reserving the grouping scheme corresponding to the combination, otherwise, screening out the grouping scheme corresponding to the combination;
step 4, for the grouping schemes corresponding to the combinations reserved in the step 3, the combined attribute labels are the attribute labels of the corresponding grouping schemes, if the number of the guarantee data in a certain grouping scheme is less than 40, the grouping scheme is deleted, for the remaining grouping schemes, a chi-square test method is adopted to carry out two-time crossed chi-square test, if the chi-square tests of two-time crossed grouping schemes both obviously do not belong to the same empirical probability distribution on a 99.9% confidence level, both the two grouping schemes are reserved; otherwise, the two grouping schemes are combined into one grouping scheme;
step 5, for all the grouping schemes obtained in the step 4, performing probability distribution fitting and probability distribution inspection on each grouping scheme to obtain the final probability distribution of the grouping scheme;
step 6, regarding the guarantee data of the service time of the fresh water truck to be predicted, if the guarantee data meets all attribute labels of a certain grouping scheme obtained in the step 4, taking the grouping scheme as an alternative grouping of the guarantee data, and thus obtaining all alternative groupings;
step 7, for each alternative group, according to preset prediction precision, performing initial prediction on the service time of the to-be-predicted water wagon to obtain the lower limit and the upper limit of the initial predicted service time of the water wagon corresponding to the alternative group;
and 8, obtaining a final prediction result according to the initial prediction fresh water truck service time lower limit and the initial prediction fresh water truck service time upper limit which correspond to all the alternative groups respectively.
2. The method for predicting the service time of the clear water vehicle at the airport according to claim 1, wherein the airport at the origin, i.e. the airport to be predicted in step 1; the planned departure time is the planned departure time of a departure place airport wheel-withdrawing gear which is declared in advance by departure flights, and the time precision is 5 minutes; the actual departure time is the actual time for departure flights to withdraw the wheel gear from the origin airport in the actual operation, and the time precision is 1 minute; the service starting time and the service ending time of the clear water vehicle are respectively the time points when the departure flight is actually running, the airport clear water vehicle is actually the time point when the departure flight starts to fill clear water and ends to fill clear water, and the time precision is 1 minute.
3. The method for predicting airport water truck service time according to claim 1, wherein the cleaning and calculating of the guarantee data in step 1 is specifically as follows:
for each piece of guarantee data, if at least one of the fresh water truck starting service time and the fresh water truck ending service time of the piece of data is in an error form, removing the piece of data; and (3) subtracting the service starting time of the clean water vehicle from the service ending time of the clean water vehicle of the data to obtain the service time of the clean water vehicle of the data, wherein the time precision of the service time of the clean water vehicle is 1 minute.
4. The method for predicting airport water truck service time according to claim 1, wherein the number of grouping schemes for ensuring data division under each combination, i.e. complexity, in step 3 is calculated by the formula:
wherein D represents the complexity of the guaranteed data partitioning calculated when m attributes are selected from the remaining 7 attributes and combined, m represents the total number of all attributes used for calculating the complexity, n is 1, …, m, anIndicates the total number of data with the n-th attribute removed, i is 1, …, an；
Attributes requiring deduplication are planned departure time and driving time, converting planned departure time with a time precision of 5 minutes into planned departure time with a precision of 1, 2, 3, 4, 6, 8, 12, or 24 hours, and converting driving time with a time precision of 1 minute into driving time with a precision of 1, 2, 3, 4, 6, 8, 12, or 24 hours.
5. The method for predicting airport water truck service time according to claim 1, wherein step 5 comprises performing probability distribution fitting and probability distribution inspection on each grouping scheme to obtain the final probability distribution of the grouping scheme, specifically:
aiming at a grouping scheme, 20 kinds of probability distribution are adopted to respectively fit the service time of the water truck, and a maximum likelihood method is used for solving each probability distribution parameter, wherein the 20 kinds of probability distribution comprise 4 types: single-parameter probability distribution, two-parameter probability distribution, three-parameter probability distribution and four-parameter probability distribution; the single parameter probability distribution includes: exponential distribution, poisson distribution, and rayleigh distribution; the two-parameter distribution includes: bernbaum sanded distribution, extremum distribution, gamma distribution, semi-normal distribution, inverse gaussian distribution, logical distribution, log-normal distribution, mid-upper distribution, normal distribution, leis distribution, and weibull distribution; the three-parameter distribution includes: a Bohr distribution, a generalized extreme distribution, a generalized pareto distribution, and a t-position scale distribution; the four parameter distribution includes: stable distribution; and (4) carrying out chi-square test on the 20 probability distribution fitting results of the grouping scheme, and selecting the probability distribution with the highest goodness of fit as the final probability distribution of the grouping scheme.
6. The method for forecasting airport fresh water vehicle service time as claimed in claim 1, wherein the calculation formula of the initial forecast in step 7 is:
wherein d represents a predicted correct probability value under a preset prediction precision, b represents a predicted lower limit of the service time of the fresh water truck, Δ c represents a preset precision of the predicted service time of the fresh water truck, b + Δ c represents a predicted upper limit of the service time of the fresh water truck, and p (x) represents a probability distribution function of the alternative grouping;
when d'bWhen the value is constantly larger than 0, the value of b is the maximum value of the definition domain minus delta c; when d'bWhen the value is constantly less than 0, the value of b is the minimum value of the definition domain; when d'bWhen the value of b is not constantly more than and not constantly less than 0, the value of b is d'bMaximum value of d when equal to 0.
7. The method for forecasting airport fresh water vehicle service time as claimed in claim 1, wherein the calculation formula of the step 8 is:
wherein, F0And F1Respectively representing the lower and upper limits of the final prediction result,
and
respectively representing the lower and upper limits of the initial prediction of the jth candidate packet, GjRespectively representing the goodness of fit of the probability distribution corresponding to the jth candidate group, and k representing the number of all the candidate groups.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110300643.8A CN113065692A (en) | 2021-03-22 | 2021-03-22 | Prediction method for service time of airport water truck |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110300643.8A CN113065692A (en) | 2021-03-22 | 2021-03-22 | Prediction method for service time of airport water truck |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113065692A true CN113065692A (en) | 2021-07-02 |
Family
ID=76562947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110300643.8A Pending CN113065692A (en) | 2021-03-22 | 2021-03-22 | Prediction method for service time of airport water truck |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113065692A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845856A (en) * | 2017-02-15 | 2017-06-13 | 民航成都信息技术有限公司 | Civil Aviation Airport ground service ensures real-time dynamic decision method |
| CN109949005A (en) * | 2019-03-07 | 2019-06-28 | 中国民用航空总局第二研究所 | A kind of airdrome scene flight ensures method, system and the terminal of operating analysis |
| US10490086B1 (en) * | 2018-10-12 | 2019-11-26 | Flightaware, Llc | System and method for collecting airport ground positional data and transmitting notifications for ground-based aircraft and other airport vehicles |
| CN111291888A (en) * | 2020-01-21 | 2020-06-16 | 西安科技大学 | Scheduling optimization method for airport special vehicles |
| CN111950910A (en) * | 2020-08-13 | 2020-11-17 | 青岛民航凯亚***集成有限公司 | Airport guarantee vehicle task scheduling method based on DBSCAN-GA |
| CN112070355A (en) * | 2020-08-05 | 2020-12-11 | 北京交通大学 | Distribution scheduling method for airport ferry vehicle |
| CN112330186A (en) * | 2020-11-18 | 2021-02-05 | 杨媛媛 | Method for evaluating ground operation guarantee capability |
-
2021
- 2021-03-22 CN CN202110300643.8A patent/CN113065692A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845856A (en) * | 2017-02-15 | 2017-06-13 | 民航成都信息技术有限公司 | Civil Aviation Airport ground service ensures real-time dynamic decision method |
| US10490086B1 (en) * | 2018-10-12 | 2019-11-26 | Flightaware, Llc | System and method for collecting airport ground positional data and transmitting notifications for ground-based aircraft and other airport vehicles |
| CN109949005A (en) * | 2019-03-07 | 2019-06-28 | 中国民用航空总局第二研究所 | A kind of airdrome scene flight ensures method, system and the terminal of operating analysis |
| CN111291888A (en) * | 2020-01-21 | 2020-06-16 | 西安科技大学 | Scheduling optimization method for airport special vehicles |
| CN112070355A (en) * | 2020-08-05 | 2020-12-11 | 北京交通大学 | Distribution scheduling method for airport ferry vehicle |
| CN111950910A (en) * | 2020-08-13 | 2020-11-17 | 青岛民航凯亚***集成有限公司 | Airport guarantee vehicle task scheduling method based on DBSCAN-GA |
| CN112330186A (en) * | 2020-11-18 | 2021-02-05 | 杨媛媛 | Method for evaluating ground operation guarantee capability |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105573329A (en) | Attitude and orbit control data analysis method based on decision tree | |
| CN105829982B (en) | For predicting that the method and system of the attended operation of aero-engine will be applied to | |
| CN111652427A (en) | Flight arrival time prediction method and system based on data mining analysis | |
| CN103646167A (en) | Satellite abnormal condition detection system based on telemeasuring data | |
| CN103646138A (en) | Time terminated acceleration acceptance sampling test optimum design method based on Bayesian theory | |
| CN104573873B (en) | A kind of terminal passenger departure's method for predicting based on fuzzy decision-tree | |
| CN110083801B (en) | OD travel time reliability estimation method and system based on robust statistics | |
| CN110738523B (en) | Maintenance order quantity prediction method and device | |
| CN110570693A (en) | Flight operation time prediction method based on reliability | |
| CN111192090A (en) | Seat allocation method and device for flight, storage medium and electronic equipment | |
| CN106971051A (en) | A kind of flight ensures service time evaluation method and system | |
| Kantor et al. | Theoretical foundation for a learning rate budget | |
| CN105825013A (en) | Aviation command guaranty optimization method based on discrete event system | |
| CN113723637B (en) | Economic repair level analysis method and system for ship repair system | |
| CN113065692A (en) | Prediction method for service time of airport water truck | |
| CN112700065B (en) | Business process completion time interval prediction method and system based on deep learning | |
| CN106355541A (en) | Method for recognizing difficulty in taking taxi on basis of GPS data | |
| Han et al. | A bi-objective optimization of airport ferry vehicle scheduling based on heuristic algorithm: A real data case study | |
| CN115600807A (en) | Method for evaluating railway passenger arrival check efficiency of public transportation operation | |
| CN114884813A (en) | Network architecture determination method and device, electronic equipment and storage medium | |
| CN107222328A (en) | Predict the method and device of saturation flex point | |
| CN111898895A (en) | Vehicle quality evaluation method and system based on big data fusion | |
| CN112529416A (en) | Construction method of operation efficiency estimation model of transportation system | |
| CN113872200B (en) | Medium-voltage distribution network power failure event diagnosis and identification method, system and storage medium | |
| CN110362437B (en) | Automatic method for embedded equipment defect tracking based on deep learning |
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 |