A new branch and bound method for solving tourism itinerary planning problems

Through the mixed integer linear planning model and the new branch bounding algorithm to optimize tourism route planning, the problem of failure to consider the queuing of service resources in the existing technology is solved, and the global optimal solution is achieved, which improves tourism service efficiency and tourist satisfaction.

CN114511136BActive Publication Date: 2024-11-26WUXI YIWEI TECHNOLOGY CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202111660109.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-31
Publication Date
2024-11-26
Estimated Expiration
2041-12-31

AI Technical Summary

Technical Problem

The existing tourism itinerary planning algorithm fails to effectively consider the queueing of service resources, resulting in tourists waiting in line at scenic spots, which will seriously decline after a given total travel time or service satisfaction, affecting the quality of tourism and tourist satisfaction.

Method used

The mixed integer linear planning model is used combined with a new branch bounding algorithm, taking into account the service capabilities of scenic spots, optimizing tourism route planning, ensuring the maximum total utility of tourists, reducing queue time, and improving satisfaction.

Benefits of technology

Through the global optimal solution of tourism route planning, the operational efficiency of the tourism service system is improved, the queue time of tourists is reduced, the satisfaction of tourists is improved, and the problems of tourism resource bottlenecks and tourist satisfaction in the existing technology are solved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114511136B_ABST
    Figure CN114511136B_ABST
Patent Text Reader

Abstract

The present invention discloses a new branch and bound method for solving the tourism itinerary planning problem, step S1: collect information of all tourists before the algorithm starts; step S2: determine the parameters required for solving the optimization model; step S3: on the basis of considering the resource limitations of tourism resources, establish an optimization model of mixed integer linear programming to calculate the planning scheme; step S4: for the optimization model, use the new branch and bound algorithm to solve, and finally obtain the tourism route planning scheme with the maximum total utility of tourists, which is the global optimal solution. The present invention considers the tourism route planning problem in my country from the perspective of global tourism, which can improve the local congestion problem that may occur in traditional tourism, and help improve the efficiency of the tourism service system and improve the tourism satisfaction of tourists.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of tourism systems, and in particular to a new branch-and-bound method for solving tourism itinerary planning problems. Background Art

[0002] At present, the travel itinerary planning of tourists mainly depends on the customers' own choices, which makes some tourist attractions seriously congested. For example, during the Golden Week, traditional hot tourist attractions are overcrowded, and even cause tourists to stay, resulting in low tourism quality and low tourist satisfaction, which has a strong negative impact on the society and environment of the destination. Therefore, in the context of vigorously promoting global tourism, it is of great practical significance to use operations planning algorithms to recommend travel itineraries for tourists, coordinate tourism resource bottlenecks, maximize tourist satisfaction, and improve the utilization rate of tourism resources from a global perspective.

[0003] The existing travel itinerary planning algorithms do not take the queuing situation of service resources into consideration. Therefore, the obtained travel itinerary planning schemes are not ideal in practice. This may cause tourists to exceed the given total travel time or have a serious decline in service satisfaction due to waiting in line at attractions.

[0004] The present invention solves the problem of tourist route planning from the perspective of the service system, helps to improve the operational efficiency of the tourist service system, reduces the queuing time of tourists, and improves the satisfaction of tourists. Summary of the invention

[0005] In view of the existing problems, the present invention provides a new branch and bound method for solving the tourism itinerary planning problem. It considers the limited service capacity of scenic spots for route planning, establishes a mixed integer linear programming model, and designs a new branch and bound algorithm, which can obtain the global optimal solution to the tourism itinerary planning problem.

[0006] In order to solve the problems existing in the prior art, the technical solution of the present invention is:

[0007] A new branch and bound method for solving the travel itinerary planning problem includes the following steps:

[0008] Step S1: Collecting information of all tourists before the planning start time, wherein the collected tourist information at least includes the total travel time of each tourist, the utility that each tourist can obtain after receiving the service of each scenic spot, etc.;

[0009] Step S2: Before the algorithm runs, determine the parameters required for the algorithm calculation; these mainly include the service time required for each scenic spot to serve a tourist, the resource constraints of each scenic spot, that is, the number of tourists that each scenic spot can serve at the same time, and the travel time between any two scenic spots;

[0010] Step S3: Taking the total utility of all tourists as the objective function, a mixed integer linear programming model is established;

[0011] Step S4: A new branch and bound algorithm is designed to solve the above optimization model, and finally a travel route planning scheme with the maximum total utility of tourists is obtained;

[0012] In S3, the mixed integer linear programming model is further defined as follows:

[0013] Step S31: Define 0-1 decision variable x ijk (1 when tourist i chooses to go from attraction k to attraction j, otherwise 0), 0-1 decision variable z ijm (When tourist i occupies resource n when visiting attraction j, it is 1, otherwise it is 0), 0-1 decision variable q iljm (For resource m at attraction j, it is 1 when tourist i is served before tourist j, otherwise it is 0); s ij u is the starting service time of tourist i at scenic spot j; ij It is an auxiliary variable used to prevent loops; tt jk is the travel time between attractions j and k, R ik is the utility of tourist i visiting attraction k; sc j is the total load of tourism resources in scenic spot j; st j The service time for attraction k; tl i is the total time spent by tourist i; N and H are the total number of tourists and the overview of the attractions.

[0014] Step S32: Establish the following total utility maximization optimization model (this model is a linear model and can be solved by a general solver such as CPLEX):

[0015]

[0016]

[0017] The new branch and bound algorithm included in step S4 further includes the following steps:

[0018] Step S41: construct the data structure of the algorithm, as follows:

[0019] Each solution uses N+H partial arrangements to represent the connection between N tourists and H attractions. Each tourist starts from the starting point (attraction 0) and ends at the end point (attraction H+1). Each tourist's travel path is represented by T k (k=1,2,...,N) represents a sequence of scenic spots. Due to resource constraints, each scenic spot must serve tourists in a certain order, which can be represented by a tourist sequence, namely R l(l=1,2,...,H). When a tourist visits a scenic spot, he occupies a position in the tourist path and tourist arrangement. For example, given N=2 and H=6, Figure 1 A solution representation is given. T1 and T2 represent the travel paths of two tourists respectively. Because both tourists visit scenic spot 5 or scenic spot 2, there will be a queue at scenic spot 5 or scenic spot 2, and there is also a queue order. R5 and R2 represent the arrangement of tourists at scenic spot 5 or scenic spot 2, that is, the queue order. Figure 1 In the figure, 2 in T2 represents scenic spot 2, and 2 in R2 represents tourist 2. When tourist 2 visits scenic spot 5, he is in the second position in the tourist path (T2) and the first position in the tourist arrangement (R5).

[0020] Step S42: The branching process of the algorithm is as follows:

[0021] The new branch and bound algorithm of the present invention plans a travel route for each tourist in turn, and the planning order has no effect on the final result because all possible routes will be tested. The new branch and bound algorithm assigns a travel path to each tourist according to the basic principle of first come first served. In each iteration, the algorithm selects a scenic spot from the scenic spots that the current tourist has not visited according to the sequence number from small to large. Tourists can choose to visit the selected scenic spot in this iteration and generate a branch. If the current tourist chooses to visit the selected scenic spot, the algorithm enters the next iteration; otherwise, the next scenic spot is selected and a new branch is generated. If the current tourist chooses not to visit the last scenic spot or has visited all scenic spots, the tourist visits the end point, and the algorithm starts to assign a route to the next tourist. For example, given N=2 and H=2, Figure 2 The figure shows the process of the algorithm developing a route for tourist 1. The maximum number of branches for tourist 1 is three to have a complete path. When implementing the algorithm, recursion can be used to implement branches.

[0022] Step S43: The delimitation process of the algorithm is as follows:

[0023] The efficiency of the branch and bound algorithm mainly depends on the efficiency of bounding, i.e. pruning. The new branch and bound algorithm of the present invention adopts a different idea from the traditional branch and bound algorithm, as follows: after the path of the pth tourist has been assigned, the algorithm calculates the following linear integer programming model (this model is a linear model and can be solved by a general solver such as CPLEX) to terminate inefficient branches (i.e. planning subproblems) during the branching process.

[0024]

[0025]

[0026] in It indicates the value of the decision variable corresponding to the time when the first p tourists have been assigned a plan.

[0027] Solve the above mixed linear integer programming model. If the solver has no solution, or the current If the corresponding upper bound is smaller than the current best solution, or all tourists have been assigned, the algorithm backtracks; otherwise, the algorithm continues to assign a path for the next tourist.

[0028] Compared with the prior art, the present invention solves the problem of travel route planning for tourists from the perspective of comprehensive consideration of global tourism. By adopting the technical solution of the present invention, travel route planning can be optimized for tourists, which helps to improve the operational efficiency of the tourism service system, reduce tourists' waiting time, and improve tourists' satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram of the data structure of the algorithm solution of the present invention;

[0030] Figure 2 Schematic diagram of the branching process of the algorithm of the present invention. DETAILED DESCRIPTION

[0031] The specific implementation modes of the present invention are described in detail below with reference to the accompanying drawings.

[0032] See also Figure 1 and Figure 2 This paper proposes a new branch and bound method for solving the tourism itinerary planning problem, which includes the following steps:

[0033] Step S1: Collecting information of all tourists before the planning start time, wherein the collected tourist information at least includes the total travel time of each tourist, the utility that each tourist can obtain after receiving the service of each scenic spot, etc.;

[0034] Step S2: Before the algorithm runs, determine the parameters required for the algorithm calculation; these mainly include the service time required for each scenic spot to serve a tourist, the resource constraints of each scenic spot, that is, the number of tourists that each scenic spot can serve at the same time, and the travel time between any two scenic spots;

[0035] Step S3: Taking the total utility of all tourists as the objective function, a mixed integer linear programming model is established;

[0036] Step S4: A new branch and bound algorithm is designed to solve the above optimization model, and finally a travel route planning scheme with the maximum total utility of tourists is obtained;

[0037] In S3, the mixed integer linear programming model is further defined as follows:

[0038] Step S31: Define 0-1 decision variable xijk (1 when tourist i chooses to go from attraction k to attraction j, otherwise 0), 0-1 decision variable z ijm (When tourist i occupies resource n when visiting attraction j, it is 1, otherwise it is 0), 0-1 decision variable q iljm (For resource m at attraction j, it is 1 when tourist i is served before tourist j, otherwise it is 0); s ij u is the starting service time of tourist i at scenic spot j; ij It is an auxiliary variable used to prevent loops; tt jk is the travel time between attractions j and k, R ik is the utility of tourist i visiting attraction k; sc j is the total load of tourism resources in scenic spot j; st j The service time for attraction k; tl i is the total time spent by tourist i; N and H are the total number of tourists and the overview of the attractions.

[0039] Step S32: Establish the following total utility maximization optimization model (this model is a linear model and can be solved by a general solver such as CPLEX):

[0040]

[0041]

[0042] The new branch and bound algorithm included in step S4 further includes the following steps:

[0043] Step S41: construct the data structure of the algorithm, as follows:

[0044] Each solution uses N+H partial arrangements to represent the connection between N tourists and H attractions. Each tourist starts from the starting point (attraction 0) and ends at the end point (attraction H+1). Each tourist's travel path is represented by T k (k=1,2,...,N) represents a sequence of scenic spots. Due to resource constraints, each scenic spot must serve tourists in a certain order, which can be represented by a tourist sequence, namely R l (l=1,2,...,H). When a tourist visits a scenic spot, he occupies a position in the tourist path and tourist arrangement. For example, given N=2 and H=6, Figure 1 A solution representation is given. T1 and T2 represent the travel paths of two tourists respectively. Because both tourists visit scenic spot 5 or scenic spot 2, there will be a queue at scenic spot 5 or scenic spot 2, and there is also a queue order. R5 and R2 represent the arrangement of tourists at scenic spot 5 or scenic spot 2, that is, the queue order. Figure 1In the figure, 2 in T2 represents scenic spot 2, and 2 in R2 represents tourist 2. When tourist 2 visits scenic spot 5, he is in the second position in the tourist path (T2) and the first position in the tourist arrangement (R5).

[0045] Step S42: The branching process of the algorithm is as follows:

[0046] The new branch and bound algorithm of the present invention plans a travel route for each tourist in turn, and the planning order has no effect on the final result because all possible routes will be tested. The new branch and bound algorithm assigns a travel path to each tourist according to the basic principle of first come first served. In each iteration, the algorithm selects a scenic spot from the scenic spots that the current tourist has not visited according to the sequence number from small to large. Tourists can choose to visit the selected scenic spot in this iteration and generate a branch. If the current tourist chooses to visit the selected scenic spot, the algorithm enters the next iteration; otherwise, the next scenic spot is selected and a new branch is generated. If the current tourist chooses not to visit the last scenic spot or has visited all scenic spots, the tourist visits the end point, and the algorithm starts to assign a route to the next tourist. For example, given N=2 and H=2, Figure 2 The figure shows the process of the algorithm developing a route for tourist 1. The maximum number of branches for tourist 1 is three to have a complete path. When implementing the algorithm, recursion can be used to implement branches.

[0047] Step S43: The delimitation process of the algorithm is as follows:

[0048] The efficiency of the branch and bound algorithm mainly depends on the efficiency of bounding, i.e. pruning. The new branch and bound algorithm of the present invention adopts a different idea from the traditional branch and bound algorithm, as follows: after the path of the pth tourist has been assigned, the algorithm calculates the following linear integer programming model (this model is a linear model and can be solved by a general solver such as CPLEX) to terminate inefficient branches (i.e. planning subproblems) during the branching process.

[0049]

[0050]

[0051] in It indicates the value of the decision variable corresponding to the time when the first p tourists have been assigned a plan.

[0052] Solve the above mixed linear integer programming model. If the solver has no solution, or the current If the corresponding upper bound is smaller than the current best solution, or all tourists have been assigned, the algorithm backtracks; otherwise, the algorithm continues to assign a path for the next tourist.

[0053] The following verification experiment is conducted by randomly generating cases. 10 cases with 7 attractions and 10 tourists are randomly generated. The location of the attractions is randomly generated in a 100×100 grid map. The range of the utility of the attractions is [1,20], the range of the service time is [1,50], the starting point and the end point are both at the (0,0) point of the grid map, and the resource limit of the attractions is 2, tl i Generated according to the following formula:

[0054]

[0055] Where T is the sum of the distances between any two scenic spots; is a random number between 0 and 1; μ1 and μ2 are tuning parameters.

[0056] Table 1. Comparison results between the new branch-and-bound algorithm and the traditional method

[0057]

[0058] Because the tourism planning problem is NP-hard, only small-scale cases can be tested and verified here. From the calculation results in Table 1, it can be seen that the performance comparison of the new branch-and-bound algorithm of the present invention with the CPLEX solution and the existing branch-and-bound algorithm. Among them, the existing branch-and-bound algorithm refers to the branch-and-bound algorithm given in Wang et al.'s (2018) (Wang, W., Lau, HC, & Cheng, SF (2018). Exact and heuristic approaches for the multi-agent orienteering problem with capacity constraints. Paper presented at the Computational Intelligence, IEEE Symposium Series on Computational Intelligence (IEEE SSCI), Honolulu, HI, NOV 27-DEC 01, 2017.).

[0059] As can be seen from Table 1, the three exact algorithms can obtain the global optimal solution for each problem. The average improvement rate of the new branch-and-bound algorithm of the present invention relative to CPLEX is 83.53%, indicating that it is better than the direct solution of the solver. The average improvement rate of the new branch-and-bound algorithm of the present invention relative to the existing branch-and-bound algorithm is 30.79%, indicating that the new branch-and-bound algorithm of the present invention has a very high improvement rate.

[0060] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A new branch and bound method for solving the travel itinerary planning problem, characterized in that: The steps include: Step S1: Collecting information of all tourists before the planning start time, wherein the collected tourist information at least includes the total travel time of each tourist and the utility that each tourist can obtain after receiving the service of each scenic spot; Step S2: Before the algorithm runs, determine the parameters required for the algorithm calculation; these include the service time required for each scenic spot to serve a tourist, the resource constraints of each scenic spot, that is, the number of tourists that each scenic spot can serve at the same time, and the travel time between any two scenic spots; Step S3: Taking the total utility of all tourists as the objective function, a mixed integer linear programming model is established; Step S4: A new branch and bound algorithm is designed to solve the above model, and finally a travel route planning scheme with the maximum total utility of tourists is obtained; In S3, the mixed integer linear programming model is further defined as follows: Step S31: Define 0-1 decision variable x ijk , when tourist i chooses to go from attraction k to attraction j, it is 1, otherwise it is 0; 0-1 decision variable z ijm , when tourist i occupies resource m when visiting attraction j, it is 1, otherwise it is 0; 0-1 decision variable q iljm , at attraction j, for resource m, it is 1 when tourist i is served before tourist l, otherwise it is 0; s ij u is the starting service time of tourist i at scenic spot j; ij It is an auxiliary variable used to prevent loops; tt jk is the travel time between attractions j and k, R ik is the utility of tourist i visiting attraction k; sc j is the total load of tourism resources in scenic spot j; st j The service time for attraction j; tl i is the total time spent by tourist i; N and H are the total number of tourists and attractions; Step S32: Establish the following total utility maximization optimization model: The constraints are: The new branch and bound algorithm included in step S4 further includes the following steps: Step S41: construct the data structure of the algorithm, as follows: Each solution uses N+H partial arrangements to represent the connection between N tourists and H attractions; each tourist starts from the starting point, attraction 0, and ends at the end point, attraction H+1; each tourist's travel path is represented by T k , k = 1, 2, ..., N represents, that is, it represents an arrangement of scenic spots; due to resource constraints, each scenic spot must serve tourists in a certain order, represented by a tourist arrangement, that is, R l , l = 1, 2, ..., H; when a tourist visits a scenic spot, he occupies a position in the tourist path and tourist arrangement; Step S42: The branching process of the algorithm is as follows: A travel route is planned for each tourist in turn. The planning order has no effect on the final result because all possible routes will be tested. The new branch and bound algorithm assigns a travel path to each tourist according to the basic principle of first come first served. In each iteration, the algorithm selects a scenic spot from the scenic spots that the current tourist has not visited according to the sequence number from small to large. Tourists can choose to visit the selected scenic spot in this iteration and generate branches. If the current tourist chooses to visit the selected scenic spot, the algorithm enters the next iteration; otherwise, the next scenic spot is selected and a new branch is generated. If the current tourist chooses not to visit the last scenic spot or has visited all scenic spots, the tourist visits the end point and the algorithm starts to assign a route to the next tourist. Recursion is used to implement branches. Step S43: The delimitation process of the algorithm is as follows: The efficiency of the branch and bound algorithm mainly depends on the efficiency of bounding, i.e. pruning, as follows: When the path of the pth visitor has been assigned, the algorithm calculates the following linear integer programming model to terminate inefficient branches during the branching process: The constraints are: in Indicates the decision variable value corresponding to the time when the first p tourists have been assigned a plan; Solve the above mixed linear integer programming model. If the solver has no solution, or the current If the corresponding upper bound is smaller than the current best solution, or all tourists have been assigned, the algorithm backtracks; otherwise, the algorithm continues to assign a path for the next tourist.

Citation Information

Patent Citations

  • Scheduling method of multi-stage variation hybrid flow shop with batch processor

    CN103309316A

  • Tour itinerary recommendation method and system, readable storage medium and computer equipment

    CN107301480A

  • Method and system for determining route planning of delivery and pick-up integrated consistent vehicles

    CN108932564A

  • Travel planner platform for providing quality tourism information

    US20170045365A1