Flight itinerary determination method and device, storage medium and electronic device

Abstract

The invention discloses a flight schedule determination method and device, a storage medium and an electronic device, and relates to the technical field of passenger protection. The method comprises the following steps: constructing a constraint model based on constraint conditions; building a passenger travel recovery graph based on the recovery object information and the constraint model; determining an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; and determining a target flight itinerary based on the initial flight itinerary and a variable neighborhood search strategy. Through the method, the device and the system, the problems that in related technologies, when the flight is changed unvoluntarily, an airline company cannot provide differentiated services for users under different conditions in a targeted manner, so that the travel experience of passengers is poor, and a large number of passengers choose to refund tickets are solved; and the technical problem of selecting flights of other airlines or taking other vehicles is solved.

Description

Technical Field

[0001] The present application relates to the technical field of passenger protection, and in particular to a method, device, storage medium and electronic device for determining a flight itinerary. Background Art

[0002] When a flight is involuntarily changed, the existing system will unify all affected passengers on the affected ticket segment, add an Involuntary Rescheduling Rights (IRR) logo, and then all affected passengers will have their flight options uniformly set from two days before to six days after the original segment date. Therefore, airlines have no way to provide differentiated services to users in different situations, resulting in a poor travel experience for passengers. In addition, the date range for flight rescheduling is not targeted, which will cause a large number of passengers to choose to refund their tickets and choose flights of other airlines or take other means of transportation, causing great losses to the airline's revenue.

[0003] When some passengers are unable to complete their trips as planned due to irregular flights, how to make full use of the remaining capacity of available flights to minimize delays for single-trip passengers and complete the connection between trips for multi-trip passengers is the core of the passenger trip recovery problem. The passenger trip recovery problem is a multi-objective scheduling optimization problem with complex constraints and large variables. As the number of passengers and flights increases, the solution space grows exponentially.

[0004] No effective solution has been proposed for the above problems. Summary of the invention

[0005] The main purpose of the present application is to provide a flight itinerary determination method, device, storage medium and electronic device to solve the technical problem in the related art that when a flight is involuntarily changed, the airline has no way to provide differentiated services to users in different situations, resulting in a poor travel experience for passengers, and then causing a large number of passengers to choose to refund their tickets and choose flights of other airlines or take other means of transportation.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a flight itinerary determination method is provided. The method comprises: constructing a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; constructing a passenger itinerary recovery graph based on recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has undergone involuntary changes, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; determining an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; and determining a target flight itinerary based on the initial flight itinerary and a variable neighborhood search strategy.

[0007] Furthermore, the constraints include flight space capacity constraints, passenger name record indivisibility constraints, flight connection constraints, transfer connection time constraints, departure time constraints, space sharing constraints and original space protection constraints, and constructing a constraint model based on the constraints includes: determining cost representation information based on the constraints, wherein the cost representation information includes space change costs, flight change costs, itinerary time costs, itinerary similarity costs, itinerary connection time costs and representation information of airline revenue costs; and constructing a constraint model based on the constraints and cost representation information.

[0008] Furthermore, constructing a passenger itinerary recovery graph based on the recovery object information and the constraint model includes: obtaining the recovery object information, wherein the recovery object information includes the departure place and the destination; creating an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; adjusting the initial passenger itinerary recovery graph based on the constraint model to obtain the passenger itinerary recovery graph.

[0009] Furthermore, the flight arc includes at least one cabin class arc, and different cabin class arcs have different costs.

[0010] Furthermore, determining the initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy includes: sorting the recovery objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort in descending order according to the value of the recovery objects or the type of the recovery objects; traversing the sorting result, and determining the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; determining a taboo table based on the number of flight segments and a greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection conditions; and determining the initial flight itinerary based on the taboo table and the backtracking search strategy.

[0011] Furthermore, determining a target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy includes: constructing a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules, and passenger exchange rules; performing a local search within any neighborhood structure based on the variable neighborhood search strategy and the initial flight itinerary to obtain search results, wherein the search results are used to indicate whether there is an improved solution within any neighborhood structure; and iteratively updating the initial flight itinerary based on the search results to obtain a target flight itinerary.

[0012] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a flight itinerary determination device is provided. The device comprises: a first construction module, which is used to construct a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; a second construction module, which is used to construct a passenger itinerary recovery graph based on recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has undergone involuntary changes, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; a first determination module, which is used to determine an initial flight itinerary from candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; and a second determination module, which is used to determine a target flight itinerary based on the initial flight itinerary and a variable neighborhood search strategy.

[0013] Furthermore, the constraints include flight space capacity constraints, passenger name record indivisibility constraints, flight connection constraints, transfer connection time constraints, departure time constraints, space sharing constraints and original space protection constraints. The first construction module is also used to determine cost representation information based on the constraints, wherein the cost representation information includes space change costs, flight change costs, itinerary time costs, itinerary similarity costs, itinerary connection time costs and airline revenue cost representation information; a constraint model is constructed based on the constraints and cost representation information.

[0014] Furthermore, the second construction module is also used to obtain recovery object information, wherein the recovery object information includes a departure place and a destination; create an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; adjust the initial passenger itinerary recovery graph based on the constraint model to obtain the passenger itinerary recovery graph.

[0015] Furthermore, the first determination module is also used to sort the recovery objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort in descending order according to the value of the recovery object or the type of the recovery object; traverse the sorting result, and determine the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; determine a taboo table based on the number of flight segments and a greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection conditions; determine the initial flight itinerary based on the taboo table and a backtracking search strategy.

[0016] Furthermore, the second determination module is also used to construct a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules and passenger exchange rules; perform a local search within any neighborhood structure based on a variable neighborhood search strategy and an initial flight itinerary to obtain search results, wherein the search results are used to indicate whether there is an improved solution within any neighborhood structure; and iteratively update the initial flight itinerary based on the search results to obtain a target flight itinerary.

[0017] In order to achieve the above objective, according to another aspect of the present application, a computer-readable storage medium is provided, wherein the storage medium stores a program, wherein the program executes any one of the above-mentioned flight itinerary determination methods.

[0018] In order to achieve the above-mentioned purpose, according to another aspect of the present application, an electronic device is provided, which includes one or more processors and a memory, and the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned flight itinerary determination methods.

[0019] Through this application, the following steps are adopted: construct a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of the target flight itinerary; construct a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is the object of the involuntary change of the flight, and the passenger itinerary recovery graph is used to record the candidate flight itinerary of the recovery object in a graphical manner; determine the initial flight itinerary from the candidate flight itinerary based on the passenger itinerary recovery graph and the backtracking search strategy; determine the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy, which solves the technical problem that when the flight is involuntarily changed in the related technology, the airline has no way to provide differentiated services to users in different situations, resulting in poor travel experience for passengers, and then causing a large number of passengers to choose to refund their tickets and choose flights of other airlines or take other means of transportation. Through the above steps, this application can not only quickly generate a passenger itinerary recovery plan, but also continuously optimize the plan to ensure that the airline's revenue and passenger satisfaction are maximized under the premise of meeting all constraints. This application can significantly improve the efficiency of airlines in dealing with emergencies such as flight cancellations and delays, reduce the inconvenience of passengers, and optimize resource allocation to improve overall operational efficiency. By dynamically adjusting the space capacity and timetable, we can better adapt to changes in actual operations and provide more flexible and efficient solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 is a flow chart of a method for determining a flight itinerary according to an embodiment of the present application;

[0022] Figure 2 is a passenger itinerary recovery diagram of the flight itinerary determination method in this embodiment;

[0023] Figure 3 is a flight arc decomposition diagram of the flight itinerary determination method in this embodiment;

[0024] Figure 4 is a neighborhood structure search framework diagram of the flight itinerary determination method in this embodiment;

[0025] Figure 5 is a flow chart of a variable neighborhood search algorithm of the flight itinerary determination method in this embodiment;

[0026] Figure 6 is a schematic diagram of a flight itinerary determination device provided according to an embodiment of the present application;

[0027] Figure 7 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. For example, an interface is set between this system and the relevant user or organization. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0032] The present invention is described below in conjunction with preferred implementation steps. Figure 1 is a flow chart of a method for determining a flight itinerary according to an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0033] Step S10, constructing a constraint model based on the constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary;

[0034] In the embodiment of the present invention, the constraint condition can be understood as the rules or restrictions that must be met when constructing the constraint model of the passenger itinerary recovery problem to ensure that the generated itinerary plan is both reasonable and feasible. Exemplarily, the scenario constraints of the passenger itinerary recovery problem mainly include: (1) Flight space capacity constraint: The number of passengers arranged on available flights cannot exceed the number of available seats on the flight. (2) Passenger Name Record (PNR) cannot be split constraint: A PNR may contain multiple passengers, and multiple passengers in the same PNR cannot be protected on different flights. (3) Flight connection constraint: For multi-trip passengers, the destination of the previous flight in the recovery plan must be the same as the departure point of the subsequent flight. (4) Transfer connection time constraint: For multi-trip passengers, the interval time between flights must be greater than the shortest transfer connection time specified by the airline. (5) Departure time constraint: The earliest departure time of the recovery plan must be later than the departure time of the original travel plan. (6) Space sharing constraint: The sum of the passengers arranged on the transfer flight and the passengers arranged on the regular flights included in the transfer flight cannot exceed the number of available seats on the aircraft. (7) Original cabin protection constraints: For passengers in a specific cabin class, the cabin class in the recovery plan should be as close to the cabin class of the original travel plan as possible, such as first-class passengers. No restrictions are imposed here.

[0035] The constraint model can be understood as a mathematical structure described by mathematical formulas or logical rules based on the above constraints. The constraint model defines how to generate or evaluate the feasibility and quality of flight itinerary plans based on the constraints. Exemplarily, the constraint model usually includes the following parts: Objective function: used to quantify and compare the pros and cons of different solutions, such as maximizing airline revenue and passenger satisfaction. Constraint set: a series of mathematical expressions that ensure that the solution meets all the above constraints. Variable definition: such as PNR set, flight set, cabin class set and their associated parameters and variables, such as the number of passengers, flight cabin capacity, flight time, etc. There is no restriction here.

[0036] A constraint model is constructed based on constraint conditions. The constraint model is used to guide the generation of a target flight itinerary. It can be understood as a constraint model constructed based on constraint conditions. It provides a framework for defining and evaluating flight itineraries to ensure the feasibility and optimization of the plan, which is not limited here.

[0037] Remember S f =(p1,p2,…,p k ) is the PNR arranged on flight f, then S=(S1,S2,…,S |F| ) represents a solution for passenger itinerary recovery. If S satisfies all constraints in the model, then it is a feasible solution.

[0038] Changes in passenger cabin class will affect passenger satisfaction. Upgrading will increase passenger satisfaction, while downgrading will reduce passenger satisfaction. The cost of cabin change for passenger p is Given by formula (1):

[0039]

[0040] in, is the cabin class used for the rth leg of PNRp, is the cabin class of the original travel plan r segment of PNRp, m chg is the calculation base of space variable cost, P is the PNR set, R p It is the set of flight segments included in the itinerary of PNRp.

[0041] Flight change costs The cost is determined by the number of flight changes between the passenger recovery plan and the original travel plan, and is given by formula (2):

[0042]

[0043] in, It is the flight used for the r segment of PNRp. is the flight used for the original travel plan r segment of PNRp, m fc is the calculation base of flight variable cost, and F is the set of available flights.

[0044] Travel time cost It is determined by the difference between the total time originally planned by the passenger and the total time of the recovery plan. The longer the total time, the lower the passenger satisfaction. Given by formula (3):

[0045]

[0046] in, is the original planned arrival time of PNRp, is the departure time of the original travel plan of PNRp, is the arrival time of the flight used for the r+1 segment of PNRp, is the departure time of the flight used for segment r of PNRp, m ttc It is the calculation base of total time cost of the trip.

[0047] Itinerary similarity cost It is determined by the number of flights in the recovery plan that have a time difference of 6 hours greater than that of the original plan. Given by formula (4):

[0048]

[0049] in, t o is the original planned departure time, t d is the original planned arrival time, t s is the departure time of the flight, t e is the flight arrival time, m sim is the trip similarity cost calculation base, It is the original scheduled departure time of the flight used for the r segment of PNRp.

[0050] Time cost of trip connection Determined by the connecting time of passengers with multiple flight segments. Given by formula (5):

[0051]

[0052] in, is the original scheduled departure time of the flight used for the r+1 segment of PNRp, is the original scheduled arrival time of the flight used for the rth leg of PNRp, is the departure time of the flight used for the r+1 segment of PNRp, is the arrival time of the flight used for segment r of PNRp, m ct It is the basis for calculating the cost of trip connection time.

[0053] Airline revenue and cost It is determined by the fixed cost based on time and the cost of passenger cabin adjustment. Given by formula (6):

[0054]

[0055] Among them, C t It is the base amount of loss caused by flight delays per unit time. is the delay time of PNRp.

[0056] The constraint model of the passenger itinerary recovery problem is shown in formula (7):

[0057]

[0058] Among them, m p is the passenger satisfaction cost weight, m a is the airline revenue cost weight, is the number of passengers protected in class q on flight f, is the capacity of class q on flight f, is the number of passengers with PNRp for segment r scheduled on flight f, n p is the number of passengers included in PNRp, is the destination of the flight used for the r segment of PNRp, Yes, MCT is the shortest connection time. is the number of passengers who are guaranteed seats in class q on connecting flight fL, is the number of passengers protected in class q on flight f, is the capacity of class q on flight f, It is the protection mark of the cabin class of the original travel plan. spe It is a collection of cabin classes that need to be protected by the original cabin class.

[0059] The first formula in formula (7) is the optimization objective of this problem, which includes the passenger satisfaction cost and the airline revenue. The multi-objective problem is handled by assigning different weights to different objectives. The second formula represents the flight capacity constraint. The number of passengers protected on the flight cannot exceed the number of available seats on the flight. The third formula represents the PNR indivisibility constraint. Multiple passengers contained in a PNR cannot be protected on different flights. The fourth formula represents the flight connection constraint. The destination of the passenger's previous flight must be the same as the departure point of the subsequent flight. The fifth formula represents the transfer connection time constraint. The interval between the arrival time of the previous flight and the departure time of the subsequent flight must be greater than the shortest transfer connection time. The sixth formula represents the departure time constraint. The earliest departure time of the recovery plan cannot be earlier than the departure time of the original travel plan. The seventh formula represents the space sharing constraint. The sum of the passengers arranged on the transfer flight and the passengers arranged on the regular flights included in the transfer flight cannot exceed the number of available seats on the aircraft. The eighth formula represents the original space protection constraint. For passengers of a specific class, when restoring the itinerary, the original class should be protected as much as possible.

[0060] In an embodiment of the present invention, given the disturbed passenger data and available flight data, based on the itinerary of the disturbed passengers, cabin class and other factors, and on the available flights, a travel itinerary that meets the travel needs of each passenger is combined while satisfying all constraints, so as to maximize the airline's revenue and passenger satisfaction.

[0061] Step S11, constructing a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has been involuntarily changed, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner;

[0062] In the embodiment of the present invention, the recovery object information can be understood as referring to the relevant data of passengers or PNRs who need to reschedule their flight itineraries due to involuntary changes in flights (such as cancellations, delays, etc.). For example, this application takes into account the operations and regulations of the actual operation of the airline, and makes the following assumptions on the premise of meeting the operational requirements of the airline: (1) After the occurrence of an abnormal flight, the flight schedule and crew member allocation plan determined by the airline before the recovery process of the disrupted passenger itinerary will not change. (2) The recovery object in this application is the disrupted passengers, and does not include other passengers on other flights who can complete the itinerary normally. (3) The available cabin capacity of the flight is the number of available seats on the flight and is greater than the number of disrupted passengers. (4) The original travel needs of the disrupted passengers will not change. (5) The departure time of the recovery plan of the disrupted passenger must be later than the departure time of the original planned travel plan. The recovery object information may include the passenger's original flight information, PNR information, affected flight information, flight network information, etc., which are not limited here.

[0063] The passenger itinerary recovery graph can be understood as a graphical data structure that is used to represent all possible flight itinerary arrangements and their interrelationships, so as to facilitate the understanding and optimization of the passenger itinerary recovery problem. Figure 2 is a passenger itinerary recovery diagram of the flight itinerary determination method in this embodiment, such as Figure 2 As shown, taking two disturbed PNRs as an example, the passenger itinerary recovery graph is used to show the passenger itinerary recovery process, and the data is preprocessed by constructing a passenger itinerary recovery graph. The graph contains all feasible solutions, and subsequent algorithms are solved based on the passenger itinerary recovery graph. The passenger itinerary recovery graph mainly contains two elements: nodes and arcs. The nodes mainly include start nodes, end nodes, PNR nodes, and flight origin and destination nodes, among which the flight origin and destination nodes are classified according to specific locations. The arc originating from the start node indicates the beginning of the PNR itinerary, and the PNR node points to the flight origin and destination nodes to indicate the demand for the PNR to take a flight. The arcs between the origin and destination nodes of flights of different categories represent specific flights, the arcs between the origin and destination nodes of flights of the same category represent the connection arcs of the flights, and the arcs pointing to the end node represent the end of the PNR itinerary. Since the present application takes into account the impact of cabin class on the recovery plan, each flight arc can be decomposed into multiple cabin class arcs, and the costs of different cabin class arcs are different. Figure 3 is a flight arc decomposition diagram of the flight itinerary determination method in this embodiment, such as Figure 3 As shown in FIG. 1 , taking a flight arc as an example, the decomposition into cabin class arcs is introduced, wherein the cabin classes are divided into first class F, first class discount A, business class J, and business discount Y.

[0064] The construction of the passenger itinerary recovery graph based on the recovery object information and constraint model can be understood as a graph network formed by connecting nodes and edges, which contains all possible itineraries of the recovery objects and the various costs and prices incurred by these arrangements. By searching and optimizing in the graph, a flight itinerary recovery plan that satisfies all constraints and maximizes airline revenue and passenger satisfaction can be found.

[0065] In the embodiment of the present invention, a passenger itinerary recovery graph is constructed based on the recovery object information and the constraint model, which transforms the complexity and variability of the flight scheduling problem into a structured and visualized graph network, so that the algorithm can perform effective search and optimization on this basis to generate the best itinerary arrangement that meets passenger needs and airline operating rules.

[0066] Step S12, determining an initial flight itinerary from candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy;

[0067] In an embodiment of the present invention, the backtracking search strategy can be understood as starting from a PNR node in the passenger itinerary recovery graph and trying to arrange flights that meet the constraints for the PNR. When selecting flights, priority is given to options that have the least impact on passenger satisfaction and airline revenue. If, when arranging a flight for a PNR, it is found that the subsequent PNR arrangement cannot meet the constraints (for example, the seats on the flight are full, or the itinerary connection time does not meet the regulations), then "backtrack" to the previous decision point, cancel the flight arrangement for the current PNR, and try other flights. If all possible flights cannot generate a feasible recovery plan for the current PNR, the recovery plan for the PNR is added to the taboo list to avoid trying the same failed path again in subsequent searches, and continue to look for recovery plans for other PNRs, which are not restricted here.

[0068] The candidate flight itinerary can be understood as starting from the PNR node in the passenger itinerary recovery graph, and multiple flight nodes can be reached through the connected edges (flight arcs). The multiple flight nodes represent the possible flight options provided for the PNR, that is, the candidate flight itinerary, which is not restricted here.

[0069] The initial flight itinerary can be understood as the first set of flight recovery options arranged for the disrupted passengers at the beginning of the algorithm based on the backtracking search strategy and the feasible flight combinations found in the passenger itinerary recovery graph.

[0070] Determining the initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy can be understood as the backtracking search strategy is used to explore possible flight combinations in the passenger itinerary recovery graph to determine the candidate flight itineraries that both meet the constraints and are feasible. The feasible flight combinations that meet the basic constraints determined from these candidate itineraries constitute the initial flight itinerary, providing a starting point for the subsequent optimization algorithm.

[0071] In the embodiment of the present invention, the initial flight itinerary is determined from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy, which ensures that the algorithm can find a suitable recovery solution in a complex flight scheduling problem, while avoiding invalid search paths and improving the efficiency of the algorithm.

[0072] Step S13, determining the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy.

[0073] In the embodiment of the present invention, the variable neighborhood search strategy can be understood as a meta-heuristic algorithm for solving combinatorial optimization problems. Compared with the local search algorithm, the variable neighborhood search algorithm searches different neighborhoods of the initial solution by constructing multiple neighborhood structures, iteratively optimizes the target value, and avoids the situation where the simple local search algorithm is prone to fall into the local optimum. As an improved local search algorithm, the variable neighborhood search can gradually optimize the target by performing local searches in different neighborhood structures. It is based on the following two core ideas: (1) The local optimal solutions in different neighborhoods are not necessarily the same; (2) The global optimal solution is the local optimal solution of all possible neighborhoods in theory. Therefore, the variable neighborhood search mainly includes five core steps: generation of the initial solution, design of the neighborhood structure, search strategy between different neighborhood structures, local search process and stopping criteria.

[0074] The iterative optimization of variable neighborhood search starts with an initial feasible solution. For some problems with simple constraints, such as the knapsack problem, random selection or greedy methods can be used to generate initial solutions. For problems such as personnel scheduling and flight scheduling, a feasible solution can be generated by direct deferred methods. The quality of the initial solution is directly related to the efficiency of optimization. Balancing the complexity of the initial solution algorithm and the quality of the initial solution is the core issue of this stage.

[0075] The optimization process of the neighborhood structure design algorithm is to search for a better solution in the neighborhood solution space of the current solution to update the current solution, and then construct a new neighborhood solution space based on the better solution to continue searching. The neighborhood structure is generated according to different neighborhood rules. Different neighborhood structures expand the search space and avoid falling into the local optimum to a certain extent. The design strategy of the neighborhood structure determines the optimization efficiency of the algorithm.

[0076] Search strategies between different neighborhood structures Since the global optimal solution is the local optimal solution of all possible neighborhoods in theory, if you want to find the solution closest to the optimal solution, you should search all solutions in all neighborhoods in theory. In the actual design process, the design of neighborhood rules generally does not include all possible neighborhoods, but in order to obtain a better solution, all neighborhoods are generally searched in order according to the size of the neighborhood, that is, starting from N(1) until N(max) is searched. Figure 4 is a neighborhood structure search framework diagram of the flight schedule determination method in this embodiment, such as Figure 4 As shown in the figure, when a better solution than the current solution cannot be found in the current neighborhood, the search is switched to the next neighborhood, as shown by the dotted line in the figure. If a better solution is found in the current neighborhood, the current solution is updated, and the neighborhood set is reconstructed based on the better solution, and the search is continued in the first neighborhood set, as shown by the solid line in the figure.

[0077] The local search process is the process of finding the local optimal solution in a neighborhood. In many problems, the scale of the solution in the neighborhood may be very large, and it takes a lot of time to traverse each neighborhood. There are two commonly used methods: the improved solution method and the optimal improved solution method. Therefore, the improved solution method is to update the current solution with the first better solution found in the search process, and update the solution space accordingly. The optimal improved solution method is to traverse the entire neighborhood and find the optimal solution in the neighborhood.

[0078] The commonly used stopping criteria of the variable neighborhood search algorithm are mainly the following: the number of times the entire neighborhood structure is traversed, the number of iterative searches within the neighborhood structure, the number of iterations of the solution process, and the length of the running time. The general steps of the variable neighborhood search algorithm are: Step 1, set x as the initial solution, initialize the neighborhood set N k (k=1,2,…,k max ), maximum number of iterations t max , k = 1, t = 1, go to step 2. Step 2, if t is greater than or equal to t max Then go to step 6; otherwise go to step 4. Step 3, if k is greater than or equal to t max Then go to step 2; otherwise go to step 4. Step 4: Randomly select a neighborhood structure k = random(1, k), and generate a neighborhood solution set N based on the initial solution x K (x), in N K (x) is randomly selected to obtain the solution x′, and then go to step 5. Step 5: Perform a local search based on x′ to obtain the local optimal solution x″, and compare the values ​​of f(x″) and f(x). If the solution x″ is better, then x=x′, t=t+1, k=1 and go to step 2; otherwise t=t+1, k=k+1 and go to step 3. Step 6, end and output the optimal solution.

[0079] The passenger itinerary recovery problem takes passengers as the scheduling object and arranges passengers in different cabins on different flights. For solution x, a set of feasible recovery solutions can be generated by defining special rules to perturb the passenger recovery solution, that is, the neighborhood N(x) of x. The neighborhood structure design of different problems is generally different. Although the basic idea of ​​the variable neighborhood search algorithm is that the global optimal solution is the local optimal solution of all possible neighborhoods, a neighborhood structure cannot be completely random and huge in scale, which is no different from enumeration. A better design is one with an acceptable neighborhood size and a higher probability of finding a better solution.

[0080] Determining the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy can be understood as exploring more solution spaces by dynamically changing the search neighborhood to find a flight itinerary that not only satisfies the constraint model but also maximizes the airline's revenue and passenger satisfaction.

[0081] In the embodiment of the present invention, the target flight itinerary is determined based on the initial flight itinerary and the variable neighborhood search strategy. The neighborhood search can effectively explore and improve on the basis of the initial flight itinerary, and finally determine a more optimized target flight itinerary to cope with the recovery problem of involuntary flight changes.

[0082] In the embodiment of the present invention, given the disturbed passenger data and available flight data, based on the itinerary, cabin class and other factors of the disturbed passenger, and based on the available flights, a travel itinerary that meets the travel needs of each passenger is combined under the condition that all constraints are met, so as to maximize the airline's revenue and passenger satisfaction. Constructing a passenger itinerary recovery graph based on the recovery object information and constraint model is to transform the complexity and variability of the flight scheduling problem into a structured and visualized graph network, so that the algorithm can effectively search and optimize on this basis to generate the best itinerary arrangement that meets passenger needs and airline operating rules. Based on the passenger itinerary recovery graph and the backtracking search strategy, the initial flight itinerary is determined from the candidate flight itineraries, ensuring that the algorithm can find a suitable recovery solution in the complex flight scheduling problem, while avoiding invalid search paths and improving the efficiency of the algorithm. Based on the initial flight itinerary and the variable neighborhood search strategy, the target flight itinerary is determined. The neighborhood search can effectively explore and improve on the basis of the initial flight itinerary, and finally determine a more optimized target flight itinerary to deal with the recovery problem of involuntary flight changes.

[0083] Optionally, the constraint conditions include flight space capacity constraint, passenger name record indivisibility constraint, flight connection constraint, transfer connection time constraint, departure time constraint, space sharing constraint and original space protection constraint. In step S10, constructing a constraint model based on the constraint conditions includes the following execution steps:

[0084] Step S101, determining cost representation information based on constraint conditions, wherein the cost representation information includes representation information of cabin space change cost, flight change cost, itinerary time cost, itinerary similarity cost, itinerary connection time cost, and airline revenue cost;

[0085] Step S102: construct a constraint model based on constraint conditions and cost representation information.

[0086] In the embodiment of the present invention, the cost representation information can be understood as a series of numerical values ​​used to quantify the impact of different elements in the recovery plan on the airline's revenue and passenger satisfaction, reflecting the economic cost and passenger experience loss under different decisions. The cost representation information is an important parameter in the constraint model for calculating the pros and cons of the solution.

[0087] The cost of class change can be understood as the additional cost or loss of satisfaction incurred by passengers due to being re-accommodated to a different class during the flight recovery process. For example, if an economy class passenger is upgraded to business class, this may increase passenger satisfaction, but if he is downgraded to economy class, it may reduce satisfaction. The cost of class change is usually related to the difference between the original class and the new class, and may be affected by the airline's operating strategy, which is not limited here.

[0088] Flight change costs can be understood as the cost that airlines may need to bear for each flight change, such as re-ticketing, subsequent itinerary adjustments, passenger notifications, etc., which will also affect the passenger's travel experience.

[0089] The travel time cost can be understood as the cost information related to the total travel time of the passenger. The extended travel time will lead to a decrease in passenger satisfaction and may cause additional losses to the airline, such as passengers may choose to refund or change to a competing airline. The travel time cost is calculated by comparing the total time difference between the recovery plan and the original plan, which is not restricted here.

[0090] The itinerary similarity cost can be understood as a measure of the similarity between the recovery plan and the original travel plan in terms of time and route. If the recovery plan differs too much from the original plan (e.g., the time difference exceeds 6 hours), the passenger may be less satisfied with the new plan. This cost reflects the passenger's sensitivity to itinerary changes and is not restricted here.

[0091] The trip connection time cost can be understood as the cost determined by the trip connection time of multi-segment passengers. If the flight connection time is too long or too short, it may affect the travel arrangements of passengers, thereby reducing their satisfaction. The trip connection time cost is related to the waiting time for flight connection or the urgency of rushing, which is not restricted here.

[0092] The airline revenue cost can be understood as the cost determined by factors such as passenger delay time and cabin changes, including fixed cost losses caused by flight delays and variable costs caused by passenger cabin changes. The airline revenue cost is calculated by evaluating the impact of the recovery plan on the airline's revenue, with the goal of minimizing revenue losses, which is not restricted here.

[0093] Determining cost representation information based on constraints can be understood as converting specific rules and constraints in flight scheduling into quantifiable costs or prices so that the algorithm can evaluate the pros and cons of different flight recovery plans. In the passenger itinerary recovery problem, constraints may include flight capacity, PNR indivisibility, flight connection time, space sharing of transfer flights, and space level protection. Converting these constraints into cost representation information actually defines a cost value for each possible violation of the constraint, so that the additional costs or the degree of decline in passenger satisfaction that may be caused by changing the recovery plan can be evaluated. For example, if a passenger is arranged on a flight with a lower class than the original class, this will generate a space change cost, representing the loss of passenger satisfaction. Similarly, if the flight connection time in the recovery plan is too long or too short and does not meet the minimum transfer connection time specified by the airline, this will generate a trip connection time cost. Through these cost representation information, a cost function can be constructed to calculate and compare different recovery plans, which are not limited here.

[0094] The construction of the constraint model based on constraint conditions and cost representation information can be understood as the transformation of constraint conditions into information that can be understood and processed by the algorithm through mathematical modeling, while the cost representation information is used as the basis for evaluating different solutions. The construction of the constraint model lays the foundation for the subsequent generation of passenger itinerary recovery graphs and algorithm optimization, ensuring that the algorithm can find a flight recovery solution that satisfies the constraints and has the minimum cost within a given framework.

[0095] In the embodiment of the present invention, the actual operating rules and passenger demands are first converted into cost representation information, and then a constraint model is constructed based on this information to ensure that the algorithm can comply with all operational constraints during the solution process and optimize with the goal of minimizing costs or maximizing benefits. It not only takes into account the personalized needs of passengers, but also takes into account the economic benefits of airline operations. It is a decision support system that comprehensively considers multiple factors.

[0096] Optionally, in step S11, constructing a passenger itinerary recovery graph based on the recovery object information and the constraint model includes the following execution steps:

[0097] Step S111, obtaining restoration object information, wherein the restoration object information includes a departure place and a destination;

[0098] Step S112, creating an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc;

[0099] Step S113, adjusting the initial passenger itinerary recovery graph based on the constraint model to obtain a passenger itinerary recovery graph.

[0100] In the embodiment of the present invention, obtaining the recovery object information can be understood as collecting relevant information of all disturbed passengers, including but not limited to the departure and destination of the passengers. This information is the basis for constructing the passenger itinerary recovery map, which can help the algorithm understand the passenger's original itinerary and the itinerary segments that need to be adjusted after the involuntary flight change. The recovery object information also includes other details in the PNR (Passenger Name Record), such as the original departure time, original flight, cabin class, etc., which are not limited here.

[0101] Creating an initial passenger itinerary recovery graph based on the origin and destination can be understood as using the recovery object information to construct an initial passenger itinerary recovery graph. The graph is a network graph in which nodes represent specific places (airports) or states (such as PNR nodes), and edges (or arcs) represent possible flight itineraries, that is, flight paths from one place to another. In the graph, the origin and destination of each disrupted passenger are connected by at least one flight arc. These flight arcs not only represent possible flights, but can also be decomposed into different cabin class arcs to reflect the impact of cabin class on the recovery plan. The initial passenger itinerary recovery graph may not contain all constraints and cost representation information, but provides a basic framework for search and optimization in subsequent steps, which is not limited here.

[0102] Adjusting the initial passenger itinerary recovery graph based on the constraint model can be understood as adjusting the generated initial passenger itinerary recovery graph based on the constructed constraint model to ensure that all generated recovery solutions meet the constraints. The constraint model takes into account constraints in terms of flight space capacity, PNR indivisibility, flight connection time, transfer connection time, and passenger class protection. The adjustment process may involve adding or deleting flight arcs, adjusting the weights of flight arcs (for example, based on cost representation information), or constructing more refined class arcs to ensure that each path in the recovery graph represents a feasible recovery solution. The adjusted passenger itinerary recovery graph not only reflects all possible flight connections, but also takes into account various constraints, so that the recovery solutions obtained by searching in the graph can be actually applied to the adjustment of passenger itineraries without violating any operating rules of airlines or passengers, which are not restricted here.

[0103] In the embodiment of the present invention, basic passenger information is obtained, and an initial passenger itinerary recovery graph is created to provide a visual network representation of each passenger's itinerary recovery. Then, the initial graph is adjusted and optimized by applying a constraint model to ensure the feasibility and effectiveness of the algorithm when searching for recovery solutions. This series of steps provides an optimization basis for subsequent search algorithms (such as backtracking search algorithms based on taboo tables and variable neighborhood search algorithms), enabling the algorithm to find the best recovery solution that meets all constraints within a given framework.

[0104] Optionally, the flight arc in step S112 includes at least one cabin class arc, and different cabin class arcs have different costs.

[0105] In the embodiment of the present invention, the class arc can be understood as representing a flight path from one airport to another airport. Specifically, the class arc represents a specific class selection in the flight path from the departure point to the destination, such as first class, business class, economy class, etc., which is not limited here.

[0106] The flight arc includes at least one cabin class arc. The different costs of different cabin class arcs can be understood as in the optimization model of flight scheduling and passenger itinerary recovery, the flight is not only considered as a whole, but is further refined to the cabin class level. This means that the flight arc (i.e., flight path) is subdivided into multiple cabin class arcs in the model, and each class arc corresponds to a specific cabin class in the flight (such as first class, business class, economy class, etc.), which is not restricted here.

[0107] In the embodiment of the present invention, by considering the cabin class arc, the optimization algorithm can manage flight resources more meticulously, ensuring that the best recovery plan can be provided to passengers when encountering flight disruptions, cancellations or delays, while taking into account the operating costs and benefits of the airline.

[0108] Optionally, in step S12, determining the initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy includes the following execution steps:

[0109] Step S121, sorting the restoration objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort the restoration objects in descending order according to their value or type;

[0110] Step S122, traversing the sorting results, and determining the number of flight segments included in the itinerary of the current restoration object based on the passenger itinerary restoration graph;

[0111] Step S123, determining a taboo table based on the number of flight segments and the greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection condition;

[0112] Step S124, determining the initial flight itinerary based on the taboo table and the backtracking search strategy.

[0113] In the embodiment of the present invention, the preset sorting strategy can be understood as a method of sorting according to weight values ​​such as passenger value and passenger type, which is not limited here.

[0114] The number of flight segments can be understood as the number of flight paths in the passenger's original itinerary. For example, a passenger's itinerary starting from Beijing, stopping in Guangzhou and arriving in Sydney includes two flight segments, which is not restricted here.

[0115] The greedy strategy can be understood as a strategy that selects the local optimal solution at each step, hoping to reach the overall optimal solution through a series of local optimal choices. In the context of passenger travel recovery, the greedy strategy may mean choosing the earliest or closest flight to the original plan for passengers to minimize delay time or travel time cost, which is not limited here.

[0116] The taboo table can be understood as a data structure used to record flight itineraries that do not meet selection conditions (such as flight capacity restrictions, itinerary connection time requirements) in previous decisions. The taboo table is used to store recovery solutions that have been tried or performed poorly to prevent the algorithm from repeatedly considering these solutions, thereby improving search efficiency and avoiding local optimal solutions.

[0117] The recovery objects are sorted according to the preset sorting strategy, and the sorting result can be understood as sorting all passengers who need itinerary recovery according to preset rules (such as passenger value, passenger type, etc.) so as to give priority to those passengers with higher priority.

[0118] Traversing the sorting results and determining the number of segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph can be understood as sorting all passengers who need itinerary recovery according to preset rules (such as passenger value, passenger type, etc.) so as to give priority to those passengers with higher priority.

[0119] The taboo table is determined based on the number of flight segments and the greedy strategy. The taboo table is used to store flight itineraries that do not meet the selection conditions. It can be understood as using a greedy strategy (giving priority to seemingly optimal flight itineraries) to try to build a recovery plan for passengers. At the same time, all flight itineraries that do not meet the constraints (such as flight capacity, itinerary connection time, etc.) are recorded and stored in the taboo table. The taboo table is dynamically updated to help the algorithm avoid repeatedly considering these infeasible flight options and improve search efficiency.

[0120] Determining the initial flight itinerary based on the taboo table and backtracking search strategy can be understood as using a backtracking search strategy, starting from the sorted passengers, trying to build their recovery flight itinerary. If the flight selection is not feasible and needs to be adjusted, backtrack and reselect the flight according to the taboo table. This process continues until an initial recovery plan that meets basic constraints is generated for each passenger. The backtracking mechanism allows the algorithm to flexibly recover from incorrect choices, avoids the trap of local optimal solutions, and helps to find a globally feasible initial recovery plan.

[0121] For example, the existence of level arcs will result in unacceptable computing time when solving the optimal path for each passenger using k shortest path algorithm, path deletion algorithm, etc. In addition, since the disturbed passengers share available flight resources, the recovery plans of the disturbed passengers will affect each other. When the number of passengers is close to the available cabin capacity of the flight, the greedy algorithm with simple rules may not be able to find a global feasible solution. Based on this, the present application proposes a backtracking search algorithm based on a taboo table, which searches for feasible solutions according to certain priority rules. If a global feasible solution cannot be obtained, the search space is limited by the taboo table until a global feasible solution is solved. First, the disturbed passengers are sorted according to the weight values ​​such as passenger value and passenger type, and recovery plans are generated in sequence. When looking for feasible solutions, the best flight for the current passenger is given priority until it exceeds the range of the recommended protected cabin space for the passenger. If a global feasible solution cannot be generated, the recovery plan of the current disturbed passenger is deleted and added to the taboo table, and other feasible solutions are backtracked until the recovery plans of all passengers are globally feasible or the problem has no solution.

[0122] Specifically, the backtracking search algorithm based on the taboo table first sorts the PNR set in descending order according to the weights of passenger value, type, etc., so that high-value or priority type passengers can be given priority. The available flights are sorted in descending order by departure time to ensure that flights with earlier times are given priority during the search, which may help reduce flight delays. Initialize the counter (n) (the size of the PNR set), (i) (the index of the PNR currently being processed), and set the taboo table to empty. The taboo table is used to record specific recovery solutions that have been tried but failed to successfully build a global feasible solution to avoid algorithm loops or falling into local optimality. Then, all PNRs in the PNR set are traversed to generate a recovery solution for each PNR. According to the number of flight segments (k) included in its itinerary, traverse each flight segment and try to build a recovery solution. According to the number of flight segments (k) included in its itinerary, traverse each flight segment and try to build a recovery solution. Based on the greedy strategy, find the optimal recovery solution (s) for the current flight segment of the PNR, which may involve selecting the most suitable flight. If a recovery solution is found, update the available space capacity of the flight to reflect the situation where the passenger is arranged on the flight.

[0123] In the embodiment of the present invention, the preset sorting strategy is used to give priority to high-value or specific types of passengers, the number of flight segments reflects the complexity of the passenger's itinerary, the greedy strategy attempts to make a local optimal choice at each step, the taboo table records infeasible flight itineraries, and is used to guide subsequent searches to avoid invalid calculations, while the backtracking search strategy allows the algorithm to return to the previous step when faced with an infeasible solution, and find a feasible initial flight itinerary by eliminating known bad choices. The combination of these strategies and data structures ensures that an initial recovery plan that meets basic constraints can be efficiently generated under limited resources, providing a starting point for subsequent optimization algorithms.

[0124] Optionally, in step S13, determining the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy includes the following execution steps:

[0125] Step S131, constructing a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules, and passenger exchange rules;

[0126] Step S132, performing a local search in any neighborhood structure based on the variable neighborhood search strategy and the initial flight itinerary to obtain a search result, wherein the search result is used to indicate whether there is an improved solution in any neighborhood structure;

[0127] Step S133, iteratively updating the initial flight itinerary based on the search results to obtain the target flight itinerary.

[0128] In the embodiment of the present invention, a neighborhood structure is constructed based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules and passenger exchange rules. It can be understood that for a PNR set P with the same departure and destination of a certain flight segment, an available flight f on the flight segment is randomly selected, and the PNR set arranged on the flight is P s , the PNR set of other available flights on this segment is Po, that is, P s =P-Po. Three different neighborhood rules are defined, the neighborhood set N(x) = {N1(x), N2(x), N3(x)}, and the neighborhood rules are defined as follows: (1) Passenger insertion rule: randomly select PNRp from the Po set and place it on flight f, check the itinerary connection constraints, flight capacity constraints, and cabin class constraints, and generate a recovery plan if the constraints are met, update the status of flight f and add p to P s (2) Passenger deletion rule: From P sRandomly select PNRp from the set, schedule it to other flights, check whether the relevant constraints are satisfied, if so, generate a recovery plan, update the status of the relevant flights, and add p to the Po set. (3) Passenger exchange rule: Randomly select two different PNRp1 and PNRp2 from the P set, exchange the recovery plans of the two PNRs and check whether the relevant constraints are satisfied. If so, generate a recovery plan and update the status of the relevant flights.

[0129] Based on the variable neighborhood search strategy and the initial flight itinerary, a local search is performed within any neighborhood structure. The search results can be understood as, after completing the construction of the initial flight itinerary, using the variable neighborhood search strategy to locally optimize the current recovery plan, starting from the current recovery plan, searching the adjacent solution space according to the predefined neighborhood rules, and looking for a better recovery plan that can improve the overall objective function value (such as reducing the total delay time, increasing airline revenue, and improving passenger satisfaction, etc.).

[0130] Figure 5 is a flow chart of a variable neighborhood search algorithm of the flight schedule determination method in this embodiment, such as Figure 5 As shown, in the process of searching in the neighborhood, the time cost of traversing all neighborhoods is extremely high, which is no different from enumeration. In order to obtain a relatively good optimization effect within a limited time, this application adopts the optimal improved solution algorithm within the range, that is, by setting the maximum number of iterations, find the optimal improved solution that can be found within the number of iterations. For neighborhood jumps, this application adopts the principle of priority neighborhood jumps, that is, first construct a neighborhood solution set based on the current solution and search for an improved solution. If a better solution than the current solution is found, the neighborhood solution set is constructed based on the new solution and continues to search according to the same neighborhood rules. When a better solution cannot be searched, the next neighborhood rule is used to construct the neighborhood and search until all neighborhood structures are traversed. At the same time, in order to avoid falling into a local optimum, a random neighborhood solution is used as the starting point for the search.

[0131] Iteratively updating the initial flight itinerary based on the search results to obtain the target flight itinerary can be understood as updating the current initial flight itinerary based on these better results of the local search, and the final flight itinerary is the target flight itinerary.

[0132] In the embodiment of the present invention, the iterative update based on the variable neighborhood search strategy is to gradually improve the initial flight itinerary through continuous local search and solution optimization, and finally obtain a better target flight itinerary. This itinerary can better cope with the challenges in flight scheduling and provide better passenger protection and resource utilization solutions while satisfying the constraints.

[0133] In the embodiment of the present invention, the specific constraints of the problem are first summarized, and a constraint model of the passenger itinerary recovery problem is constructed. The model aims to maximize airline revenue and passenger satisfaction, and takes into account constraints such as passenger multi-trip connection, transfer flight space sharing, and PNR indivisibility. Then a passenger itinerary recovery graph is established to represent the process of passenger itinerary recovery, and the graph contains all feasible solutions. Then a backtracking search algorithm based on a taboo table is proposed to generate an initial feasible solution. Then an optimization algorithm based on variable neighborhood search is proposed. Based on the initial solution generated by the backtracking search algorithm based on the taboo table, it continuously performs iterative search through the constructed three neighborhood structures of passenger insertion, passenger deletion, and passenger exchange, and uses the method of starting local search in the random neighborhood solution of the initial solution to avoid falling into the local optimal solution. Compared with the backtracking search algorithm based on the taboo table, the variable neighborhood search optimization can effectively optimize the problem goal within a limited time, indicating the solution efficiency and optimization performance of the algorithm.

[0134] The present application also provides a flight schedule determination device. It should be noted that the flight schedule determination device of the present application can be used to execute the flight schedule determination method provided in the present application. The flight schedule determination device provided in the present application is introduced below.

[0135] Figure 6 is a schematic diagram of a flight itinerary determination device 600 provided according to an embodiment of the present application. Figure 6 As shown, the device includes: a first construction module 601, a second construction module 602, a first determination module 603, and a second determination module 604.

[0136] Specifically, the first construction module 601 is used to construct a constraint model based on the constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary;

[0137] A second construction module 602 is used to construct a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has been involuntarily changed, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner;

[0138] A first determination module 603 is used to determine an initial flight itinerary from candidate flight itineraries based on a passenger itinerary recovery graph and a backtracking search strategy;

[0139] The second determination module 604 is used to determine the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy.

[0140] In summary, the flight itinerary determination device provided by the embodiment of the present application, through the first construction module 601, constructs a constraint model based on the constraint conditions, wherein the constraint model is used to guide the generation of the target flight itinerary; through the second construction module 602, constructs a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is the object of the flight involuntary change, and the passenger itinerary recovery graph is used to record the candidate flight itinerary of the recovery object in a graphical manner; through the first determination module 603, the initial flight itinerary is determined from the candidate flight itinerary based on the passenger itinerary recovery graph and the backtracking search strategy; through the second determination module 604, the target flight itinerary is determined based on the initial flight itinerary and the variable neighborhood search strategy, which solves the technical problem that when the flight is involuntarily changed in the related technology, the airline has no way to provide differentiated services to users in different situations, resulting in poor travel experience for passengers, and then causing a large number of passengers to choose to refund their tickets and choose flights of other airlines or take other means of transportation. Through the above steps, the present application can not only quickly generate a passenger itinerary recovery plan, but also continuously optimize the plan to ensure that the airline's revenue and passenger satisfaction are maximized under the premise of meeting all constraints. This application can significantly improve the efficiency of airlines in dealing with emergencies such as flight cancellations and delays, reduce the inconvenience of passengers, optimize resource allocation, and improve overall operational efficiency. By dynamically adjusting the capacity and schedule, it can better adapt to changes in actual operations and provide more flexible and efficient solutions.

[0141] Furthermore, the constraints include flight space capacity constraints, passenger name record indivisibility constraints, flight connection constraints, transfer connection time constraints, departure time constraints, space sharing constraints, and original space protection constraints. The first construction module 601 is also used to determine cost representation information based on the constraints, wherein the cost representation information includes space change costs, flight change costs, itinerary time costs, itinerary similarity costs, itinerary connection time costs, and airline revenue cost representation information; and construct a constraint model based on the constraints and cost representation information.

[0142] Furthermore, the second construction module 602 is also used to obtain recovery object information, wherein the recovery object information includes the departure place and the destination; create an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; adjust the initial passenger itinerary recovery graph based on the constraint model to obtain the passenger itinerary recovery graph.

[0143] Furthermore, the first determination module 603 is also used to sort the recovery objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort in descending order according to the value of the recovery object or the type of the recovery object; traverse the sorting result, and determine the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; determine a taboo table based on the number of flight segments and a greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection conditions; determine the initial flight itinerary based on the taboo table and the backtracking search strategy.

[0144] Furthermore, the second determination module 604 is also used to construct a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules and passenger exchange rules; perform a local search within any neighborhood structure based on a variable neighborhood search strategy and an initial flight itinerary to obtain search results, wherein the search results are used to indicate whether there is an improved solution within any neighborhood structure; and iteratively update the initial flight itinerary based on the search results to obtain a target flight itinerary.

[0145] The flight itinerary determination device includes a processor and a memory. The above-mentioned first construction module 601, second construction module 602, first determination module 603, second determination module 604, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to implement corresponding functions.

[0146] The processor includes a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the target flight itinerary is determined by adjusting the kernel parameters.

[0147] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0148] An embodiment of the present invention provides a computer-readable storage medium having a program stored thereon, and when the program is executed by a processor, the flight itinerary determination method is implemented.

[0149] An embodiment of the present invention provides a processor, which is used to run a program, wherein the flight itinerary determination method is executed when the program is run.

[0150] like Figure 7As shown, an embodiment of the present invention provides an electronic device, the device includes a processor, a memory, and a program stored in the memory and executable on the processor, and the processor implements the following steps when executing the program: constructing a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; constructing a passenger itinerary recovery graph based on recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has undergone involuntary changes, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; determining an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; and determining a target flight itinerary based on the initial flight itinerary and a variable neighborhood search strategy.

[0151] When the processor executes the program, the following steps are also implemented: the constraints include flight space capacity constraints, passenger name record indivisibility constraints, flight connection constraints, transfer connection time constraints, departure time constraints, space sharing constraints, and original space protection constraints. Building a constraint model based on the constraints includes: determining cost representation information based on the constraints, wherein the cost representation information includes space change costs, flight change costs, itinerary time costs, itinerary similarity costs, itinerary connection time costs, and representation information of airline revenue costs; building a constraint model based on the constraints and cost representation information.

[0152] When the processor executes the program, the following steps are also implemented: constructing a passenger itinerary recovery graph based on the recovery object information and the constraint model includes: obtaining the recovery object information, wherein the recovery object information includes the departure place and the destination; creating an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; adjusting the initial passenger itinerary recovery graph based on the constraint model to obtain the passenger itinerary recovery graph.

[0153] When the processor executes the program, the following steps are also implemented: the flight arc includes at least one cabin class arc, and different cabin class arcs have different costs.

[0154] When the processor executes the program, the following steps are also implemented: determining the initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy includes: sorting the recovery objects according to the preset sorting strategy to obtain the sorting results, wherein the preset sorting strategy is used to sort in descending order according to the value of the recovery objects or the type of the recovery objects; traversing the sorting results, and determining the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; determining a taboo table based on the number of flight segments and the greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection conditions; determining the initial flight itinerary based on the taboo table and the backtracking search strategy.

[0155] When the processor executes the program, the following steps are also implemented: determining the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy includes: constructing a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules, and passenger exchange rules; performing a local search within any neighborhood structure based on the variable neighborhood search strategy and the initial flight itinerary to obtain search results, wherein the search results are used to indicate whether there is an improved solution within any neighborhood structure; iteratively updating the initial flight itinerary based on the search results to obtain the target flight itinerary.

[0156] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0157] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps: constructing a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; constructing a passenger itinerary recovery graph based on recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has undergone involuntary changes, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; determining an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; and determining a target flight itinerary based on the initial flight itinerary and a variable neighborhood search strategy.

[0158] When executed on a data processing device, it is also suitable for executing an initialization program having the following method steps: constraints include flight space capacity constraints, passenger name record indivisibility constraints, flight connection constraints, transfer connection time constraints, departure time constraints, space sharing constraints, and original space protection constraints; constructing a constraint model based on the constraints includes: determining cost representation information based on the constraints, wherein the cost representation information includes space change costs, flight change costs, itinerary time costs, itinerary similarity costs, itinerary connection time costs, and representation information of airline revenue costs; constructing a constraint model based on the constraints and cost representation information.

[0159] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: constructing a passenger itinerary recovery graph based on recovery object information and a constraint model, including: obtaining recovery object information, wherein the recovery object information includes a departure place and a destination; creating an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; adjusting the initial passenger itinerary recovery graph based on the constraint model to obtain the passenger itinerary recovery graph.

[0160] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: the flight arc includes at least one cabin class arc, and the costs of different cabin class arcs are different.

[0161] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: determining an initial flight itinerary from candidate flight itineraries based on a passenger itinerary recovery graph and a backtracking search strategy, including: sorting the recovery objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort in descending order according to the value of the recovery object or the type of the recovery object; traversing the sorting result, and determining the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; determining a taboo table based on the number of flight segments and a greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection conditions; determining the initial flight itinerary based on the taboo table and the backtracking search strategy.

[0162] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: determining a target flight itinerary based on an initial flight itinerary and a variable neighborhood search strategy includes: constructing a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules, and passenger exchange rules; performing a local search within any neighborhood structure based on the variable neighborhood search strategy and the initial flight itinerary to obtain search results, wherein the search results are used to indicate whether an improved solution exists within any neighborhood structure; iteratively updating the initial flight itinerary based on the search results to obtain a target flight itinerary.

[0163] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0164] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.

[0165] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0166] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0167] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0168] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0169] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0170] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0171] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0172] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for determining a flight itinerary, characterized in that: The method comprises: Building a constraint model based on the constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; Constructing a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has been involuntarily changed, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; determining an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; The target flight itinerary is determined based on the initial flight itinerary and a variable neighborhood search strategy.

2. The method according to claim 1, characterized in that: The constraint conditions include flight space capacity constraint, passenger name record indivisibility constraint, flight connection constraint, transfer connection time constraint, departure time constraint, space sharing constraint and original space protection constraint. The constraint model constructed based on the constraint conditions includes: Determine cost representation information based on the constraint conditions, wherein the cost representation information includes representation information of cabin space change cost, flight change cost, itinerary time cost, itinerary similarity cost, itinerary connection time cost, and airline revenue cost; The constraint model is constructed based on the constraint conditions and the cost representation information.

3. The method according to claim 1, characterized in that The constructing of the passenger itinerary recovery graph based on the recovery object information and the constraint model comprises: Acquire the restoration object information, wherein the restoration object information includes a departure place and a destination; Creating an initial passenger itinerary recovery graph based on the departure place and the destination, wherein the departure place and the destination in the initial passenger itinerary recovery graph are connected by at least one flight arc; The initial passenger itinerary recovery graph is adjusted based on the constraint model to obtain the passenger itinerary recovery graph.

4. The method according to claim 3, characterized in that The flight arc includes at least one cabin class arc, and different cabin class arcs have different costs.

5. The method according to claim 1, characterized in that Determining the initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and the backtracking search strategy includes: Sorting the restoration objects according to a preset sorting strategy to obtain a sorting result, wherein the preset sorting strategy is used to sort the restoration objects in descending order according to their value or type; Traversing the sorting results, and determining the number of flight segments included in the itinerary of the current recovery object based on the passenger itinerary recovery graph; Determine a taboo table based on the number of flight segments and a greedy strategy, wherein the taboo table is used to store flight itineraries that do not meet the selection condition; The initial flight itinerary is determined based on the taboo table and a backtracking search strategy.

6. The method according to any one of claims 1 to 5, characterized in that Determining the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy includes: Building a neighborhood structure based on target domain rules, wherein the target domain rules include passenger insertion rules, passenger deletion rules, and passenger exchange rules; Performing a local search in any neighborhood structure based on the variable neighborhood search strategy and the initial flight itinerary to obtain a search result, wherein the search result is used to indicate whether an improved solution exists in any neighborhood structure; The initial flight itinerary is iteratively updated based on the search results to obtain the target flight itinerary.

7. A flight itinerary determination device, characterized in that: The device comprises: A first building module is used to build a constraint model based on constraint conditions, wherein the constraint model is used to guide the generation of a target flight itinerary; A second construction module is used to construct a passenger itinerary recovery graph based on the recovery object information and the constraint model, wherein the recovery object information is used to represent the flight itinerary plan of the recovery object, the recovery object is an object whose flight has been involuntarily changed, and the passenger itinerary recovery graph is used to record the candidate flight itineraries of the recovery object in a graphical manner; A first determination module, configured to determine an initial flight itinerary from the candidate flight itineraries based on the passenger itinerary recovery graph and a backtracking search strategy; The second determination module is used to determine the target flight itinerary based on the initial flight itinerary and the variable neighborhood search strategy.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the flight itinerary determination method according to any one of claims 1 to 6 when running on a computer or a processor.

9. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the flight itinerary determination method as claimed in any one of claims 1 to 6.

10. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the flight itinerary determination method as described in any one of claims 1 to 6.

Images