Tourism Route Selection Model for Tourism Sustainable Development Based on Improved Genetic Algorithm

Abstract

The improvement of people’s quality of life also promotes the development of tourism. The traditional travel mode is no longer suitable for the needs of modern people. How to quickly determine the optimal tourist route according to the needs of different families is the current sustainable development direction of tourism. Scientific planning of tourist routes to minimize the cost and time of tourists is very important for the improvement of tourism experience. This study employs an improved genetic algorithm (IGA) to find the best tourist route based on this requirement. With the increase of tourist attractions, routes, and demands, the traditional genetic algorithm (GA) has problems such as premature convergence and poor local search ability when planning the tourist route. The existence of these problems will affect the effect of route planning. IGA proposes three improvements to traditional GA. One is to introduce the ant colony algorithm (ACA) to initialize the parameters in the GA. This algorithm is introduced to alleviate the over-reliance of GA on initializing the population and the poor adaptability of individual populations. Second, because the crossover probability parameter in GA has such a large influence on the final solution, this study proposes an adaptive strategy for adjusting the crossover probability to improve population fitness. Third, considering the weak local search ability of GA and the problem of premature convergence, this study introduces the 2-opt optimization algorithm to improve the quality of the solution. The results of the experimental analysis confirm the effectiveness of the proposed method.

1. Introduction

The development of the national economy has led to the prosperity of tourism. The rise of national tourism is one of the reasons for the economic development of the country. The development of international tourism can increase the country’s foreign exchange income. The development of domestic tourism can return currency and stabilize the market economy. The second is to promote the prosperity and development of social culture. Tourism can promote the improvement of national quality and people’s quality of life. It also promotes cultural exchange while providing a large number of employment opportunities. The third is to affect the regional environment. Tourism can promote environmental protection. The development of tourism also has a negative effect on the environment. If the relationship between tourism and the environment is not handled properly, the environment will also develop in the direction of deterioration. In general, the development of tourism is of great significance to the nation as well as to the individual. The number of tourists and attractions is increasing every year, but people’s time and energy are very limited. How to effectively plan travel routes and maximize time and cost savings is what every traveler cares about. Tourism planning is of great significance to relevant government departments, tourism industry, related attractions, and individuals. From the standpoint of tourism development, the goal of tourism planning is to allocate and utilize all tourism resources, as well as tourism reception capacity, in a reasonable and effective manner, transportation capacity, and human, material, and financial resources that the society may provide to the tourism industry. By means of prediction and adjustment, blindness is reduced, so as to maximize the economic, social, and environmental benefits of developing tourism and make tourism sustainable. From the perspective of national and regional economic and social development, the purpose of tourism planning is to clarify the status and role of tourism in national economic and social development, develop tourism in a planned and step-by-step manner, encourage tourism to play a more prominent role in national or local economic and social development, and promote the overall prosperity of society and the regional economy.

Tourism planning has a strategic guiding significance, and it clearly puts forward the direction, scale, speed, and goal of tourism development, as well as the strategy to achieve the goal [1–3]. Its function is mainly manifested in the following aspects: 1. set goals for tourism development. The development goals of tourism planning can define the development direction of the tourism system, seek a balance between the ideal state and the reachable state, and realize the rational utilization of tourism resources and the sustainable development of tourism. This requires that tourism planning must be based on reality, objective investigation and evaluation, and the law of tourism system development, and the development history and current situation, advantages, and disadvantages of tourism in the planning area are comprehensively analyzed. 2. Reasonable allocation of resources. In order to achieve the established goals and maximize the economic, social, and ecological benefits of tourism, it is absolutely not enough to only rely on the setting of goals. It must also mobilize all positive human, material, and financial resources on the basis of social reality to achieve multiwin, social harmony and progress, and economic development. Resource is the foundation of tourism development, market is the means of modern tourism development, and benefit is the purpose of tourism planning and development. If resource conditions are ignored, the risk of competition in the tourism market will be greatly increased. The attractiveness of tourism resources is often hidden and must be explored through certain planning and development in order to highlight its uniqueness. In addition, the attractiveness of tourism resources is largely influenced by tourists’ psychology. With the continuous progress of society and the increasing demand of tourists, tourism resources must be constantly changing and new in order to maintain their lasting appeal. Therefore, the planning and development of tourism resources are very important. Through a series of steps such as resource evaluation, location analysis, market research, development forecast, and resource protection, tourism planning scientifically and reasonably determines the combination of resources and markets. It points out the direction for the development of tourism resources, plans a competitive product system, and fully taps the advantages of tourism resources.

Reference [4] proposes a method for optimizing one-way travel routes, the ultimate goal of which is to determine the optimal route that takes the least time from the starting point to the destination. Reference [5] proposes a travel route recommendation method using social photos. Travel itinerary recommendation based on social photos is mainly based on user preferences. Reference [6] proposes two mathematical models for the random time window assignment problem and finally finds the route with the shortest driving distance. Reference [7] uses the disutility method to find the optimal travel route. Reference [8] uses heuristics to optimize travel routes with the ultimate goal of getting as many hitchhikers as possible in a short period of time. Reference [9] uses a game idea to optimize and minimize the travel time for public transportation route planning. In reference [10], for the problem of urban garbage truck routes, a clustering algorithm is used to find the route with the best travel distance and travel time. Reference [11] designed the optimal tourist route in the region by considering the geographical features of the Urals, taking into account the scenic spots, routes, time, and other factors. The route optimization algorithm proposed in reference [12] is able to simulate the ideal prediction of the speed of the road segment in the network. Reference [13] combines GA and traveling salesman algorithm to propose a route optimization algorithm. The simulation implementation shows that the route based on this method shortens the UAV flight time by 26%. The above route planning problems have different application areas and use different optimization models. Compared with the route optimization method proposed before, both in terms of time consumption and cost expenditure, it has achieved fruitful results. Aiming at the tourism industry, this study proposes a travel route optimization method for tourists. The method introduces an improved GA to find the optimal travel arrangement. Traditional GAs are prone to problems such as premature convergence and poor local search ability when planning travel routes. The existence of these problems will greatly affect the optimal route planning. In order to solve the above problems, this study proposes IGA. IGA optimizes initialization parameters and crossover probability parameters. In the later stages of evolution, the 2-opt optimization algorithm is used to solve the problem of low search efficiency and long solution time. The experimental analysis results show that the IGA algorithm can get a better route planning scheme.

2. Relevant Knowledge

2.1. Multiobjective Optimization Problem

The tourist route planning problem is a typical multiobjective optimization problem. Multiple goals include less time spent, less expense, and higher tourist satisfaction. Therefore, understanding the solution of multiobjective optimization problems is beneficial to optimize the planning of tourist routes. The mathematical expression of a typical multi-objective optimization problem is as follows:

The number of objectives is m, x is the n-dimensional decision vector, and y is the objective vector of the multiobjective optimization problem. is the m subobjective functions to be optimized, expresses h inequality constraints, and the equality constraints are , a total of n.

There are two ways to solve equation (1). One approach is to reduce the multiobjective problem to a single-objective problem. The second step is to use the optimization method to find the best solution. The two methods’ solutions are as follows:

2.1.1. Multiobjective Conversion into a Single-Objective Problem

The specific methods for solving multiobjective problems in this way mainly include the weighted summation method of objective function, the objective programming method, and the ε-constraint method. The idea of the objective function weighted sum method is to multiply each objective function in the problem by a weight and then sum these objective functions. This method can solve a multiobjective problem by converting it to a single-objective problem. The formula of the objective function weighted summation method is as follows:where , , and is the weight of each subobjective function. For nondominated multiobjective optimization problems with convex function characteristics, the solution set obtained by this method can well meet the requirements. However, its shortcomings are also very obvious, and the final result obtained is closely related to the selected weight coefficient. Therefore, the selection of weight coefficients is a very important step in practical operation.

The objective programming method’s main idea is to set the expected value of each subobjective function of the problem as G and transform it into an objective optimization problem that satisfies the following equation:

By solving equation (3), a set of solutions satisfying the conditions can be finally obtained.

The main idea of the constraint method is to take any objective function in the problem as the main optimization objective and the rest as constraints. In this manner, it can be reduced to a single-objective optimization problem satisfying the following equation:

2.1.2. Optimization Method to Solve

The most common optimization methods are GA [14], particle swarm algorithm [15], simulated annealing algorithm [16], and so on. This study relies heavily on the GA to optimize the tourist route. GA is a technique for solving single-objective optimization problems. The population is the target of the GA operation.

The GA algorithm starts from the initial population, selects excellent individuals from the population to enter the next generation population through certain rules, and then performs individual selection, crossover, mutation, and other operations on the population to obtain the optimal solution. Figure 1 shows the flowchart of GA.

Figure 1 

GA flowchart.

2.2. Route Planning Technology

Route planning technology refers to a method to help users plan routes and is divided into two categories: automatic planning and interactive planning. Automatic planning is to use the path optimization algorithm to automatically plan the optimal route under the target for the user. Interactive planning is a method for users to design a complete route through interactive functions, selecting destinations and passing places.

2.2.1. Automatic Planning

Automatic route planning is an NP-hard problem of path optimization [17]. For NP-hard problems, optimization algorithms such as heuristic algorithms are usually used to solve them. Heuristic algorithms [18, 19] are divided into two types: constructive heuristics and metaheuristics. Construction heuristics use incremental methods to iteratively add nodes until a complete solution is generated, such as greedy-based methods [20], graph construction-based methods [21], or region partition-based methods [22]. Such algorithms usually have high efficiency. However, due to structural limitations, it is easy to fall into a local optimum, and the quality of the solution is often not high enough. The metaheuristic algorithm expands the search range on this basis and uses iterative means to continuously optimize to obtain better solutions. But the time and space costs it takes are often unacceptable. Common metaheuristic algorithms include GA [23], bee colony algorithm [24], water drop algorithm, etc. [25].

2.2.2. Interactive Planning

Interactive planning refers to users interacting with online maps to complete interactive tasks related to route planning. The design of the map interaction is centered on the task flow. The process is shown in Figure 2. First, data such as time, space, and calculated values are selected. Statistics are then performed and each piece of data is analyzed. Finally, the map is drawn and displayed to the user through the client.

Figure 2 

Schematic diagram of interaction design.

Figure 3 shows the flow of using map interaction to complete route planning tasks. Users first retrieve key metrics, such as detailed addresses. The system determines the relationship of each indicator according to the user’s choice, then queries the corresponding indicators from the symbol library, makes statistics, integrates the obtained data, and encapsulates them into different layer files. Finally, the map is displayed to the user through dynamic rendering. This complete process requires repeated iterations in the process of the user planning the route until the user completes the route planning.

Figure 3 

Interaction process.

3. Improved Genetic Algorithm

When the scale of the tourist route becomes larger and the demand is higher, the traditional GA is prone to problems such as premature convergence and poor local search ability when planning the tourist route. The proposed IGA can avoid these problems. The flowchart of the IGA algorithm is shown in Figure 4.

Figure 4 

IGA flowchart.

3.1. Individual Encoding and Fitness Function

Coding operations are phenotype-to-genotype mappings. Combined with the characteristics of route planning, the path coding method is adopted, and the natural numbers of 1, 2,…,i, are used to identify the scenic spots. They are connected into a string in the order, in which the routes arrive. The fitness function used in GA has an effect on the algorithm’s convergence speed and the search for the optimal solution. The objective function of the tourism planning model is to minimize it, and the reciprocal of the objective function can be used directly as the fitness function, as shown in the following formula:

The larger the fitness function, the better the chromosome individual is, resulting in a shorter total path length.

3.2. Operators of IGA

3.2.1. Initializing the Population

The average quality of population individuals is low due to the poor fitness of GA initialized population individuals and the algorithm’s poor convergence ability. ACA is used to optimize population initialization. Therefore, the IGA can maintain a high fitness initially, ensuring that the algorithm operates at a high convergence speed. The algorithm steps are described as follows: Step 1: Initialize the parameters of the ACA. The parameters that need to be initialized are population size, maximum number of iterations, number of ant colonies, pheromone strength, information heuristic factor, and expectation heuristic factor. Step 2: Solve the GA initial solution by ACA.(1)Read the city-related data information, and calculate the mutual distance d_ij between cities i and j.(2)Initialize the visit taboo table of each ant, and each ant selects the next city according to the transition probability P. Add unvisited cities to the traversal path. Update the taboo table until all ants search to complete a search loop.(3)Update the global pheromone matrix.(4)Record the current solution of each ant, record the optimal solution, and output it. Step 3: The output result of the ACA is used as the initial population of the GA.

3.2.2. Selection Operator

The expected value method can be used to solve the TSP problem. The core idea is as follows:(1)Calculate the expected number of each chromosome individual in the population that can be inherited to the next generation. The calculation formula is as follows: where s_i is the fitness value of individual i, i = 1,2, …, n.(2)If a chromosome individual is selected to participate in the next-stage crossover operation, its M_i in the next generation is subtracted by 0.5. If not selected, M_i is subtracted by 1.(3)When the expected survival value M_i of a chromosome individual is less than zero, it will be eliminated and will not continue to participate in the selection operation.

3.2.3. Crossover Operator

In this study, the sequential crossover operator (OX) is used to complete the crossover operation. The population difference in the early stage of GA operation is relatively large. To help the algorithm find the optimal solution area and speed up the generation of new individuals, the crossover probability should be set to a high value. With algorithm advancement, it is necessary to set a small crossover probability, reduce search speed, and improve algorithm accuracy for the local optimal solution. But too large crossover probability will make GA lose its characteristics. Too small search probability makes the GA performance extremely poor.

This study introduces adaptive strategies to GA. The lower the crossover probability is set, and the better individuals in the population are retained, the higher the individual fitness value. If the fitness value of the population’s individuals is low, the crossover probability is increased. Individuals usually set a higher crossover probability in the early stages of population evolution to accelerate the rate of new individuals. A lower crossover probability is set later in the algorithm to improve the algorithm’s accuracy. The formula for calculating the adaptive crossover probability is as follows:where is the group’s maximum fitness value. is the group’s average fitness value. is the greater fitness value of the two individuals involved in the cross, which is usually a constant. c₁ and c₂ are between 0 and 1.

3.2.4. Mutation Operator

In this study, the reverse mutation operator is used to complete the mutation operation. The basic idea of reverse mutation is to randomly select the interception point on the chromosomal coding string and reverse the intercepted gene sequence according to the reversal probability. Reverse mutation is a special form of basic mutation.

3.3. IGA Execution Steps

 Step 1: Algorithm Initialization. Initialize various parameters of ACA and GA respectively. Step 2: The ACA strategy is used to initialize the population. Step 3: According to the objective function, equation (5) is used to calculate the individual fitness fi in the population. Step 4: Selection Operation. Use the expected value method to select excellent individuals to inherit to the next generation. Step 5: Crossover Operation. The adaptive crossover equation (7) is used to calculate the crossover probability Pc of the individual and generate a new individual. Step 6: Mutation Operation. Use the reverse mutation method to calculate the mutation probability Pm of the individual, select the individual according to the mutation probability, and complete the mutation operation. Step 7: The 2-opt local search operator is used to optimize the local search. Step 8: The algorithm terminates. It is judged whether the termination condition is met, if so, go to step 9; if not, go to step 3. Step 9: Output the optimal path.

4. IGA-Based Tourism Route Planning Experiment

4.1. Tourism Problem Description

Tourism route selection refers to the route from the origin to the destination that satisfies the relevant constraints. The constraints here usually refer to travel time, cost, etc. Assuming that there are 4 scenic spots in the process from the starting point x_s to the ending point x_t, the tourist route is shown in Figure 5.

Figure 5 

Tourist itinerary.

In order to select the optimal route to meet customer needs from the routes shown in Figure 6, it is necessary to establish a mathematical model to solve. In this study, IGA is used to establish a multiobjective-based travel route selection model. The data of each scenic spot are used as the basic data, and the objective function and constraints are given, and finally, a sequence of scenic spots that meets the requirements is generated. The solution process of the optimal route is shown in Figure 6.

Figure 6 

The optimal tourist route solution process.

4.2. Tourism Route Selection Model

Assuming that D_ij represents the distance between p_i and p_j scenic spots, and the coordinates of each scenic spot are described by latitude and longitude. The distance between attractions is in kilometers. p_i and p_ij are defined as follows:

Travel itinerary set . For example, means the starting point is p₁, the key point is p₉, and it passes through p₃, p₄, and p₆ on the way. The travel cost TE includes travel expenses, scenic spot tickets, and dining expenses. The toll L (p_ij) is the largest expenditure part. In the process of self-driving tour, the calculation of toll is calculated by multiplying the oil price per kilometer by the kilometers. For example, the distance between attractions p_i and p_j is D_ij, and the gas price per kilometer is 0.7 yuan. Then, the cost . The maximum travel expenditure is TE_max. The total travel time TT includes the travel time and the time spent on sightseeing. The time spent on sightseeing is ST, and the time spent on the way is LT (p_ij). The maximum travel time is TT_max.

Usually tourists have multiple requirements for the choice of tourist routes, such as the shortest time requirement and the least cost. In order to meet the needs of tourists as much as possible, the research needs to use a multiobjective constraint tourist route planning model. The main objectives of this study are cost constraints and time constraints. The details of each constraint are described as follows:

4.2.1. Travel Cost Constraints

where L_ij represents the toll from scenic spot p_i to scenic spot p_j. The total fee must be less than the given maximum fee TE_max:

Equation (10) is used to judge whether the scenic spot p_i is a scenic spot on the optimal route. If p_fi = 1, then TE_i means that the travel cost of the scenic spot p_i is the sum of the travel cost. If P_fi = 0, then it means that the scenic spot p_i is not a subset on the optimal path. TE_i indicates that the current travel cost is less than the travel cost cap set by the tourists themselves:

Constraints (11) and (12) indicate that each attraction is different, and each attraction can only be played once. After the scenic spot p_i in the optimal route, there is only one scenic spot p_j followed. Usually the total amount of expenses for tourists to travel is limited. In order to make tourists as happy as possible within the limited expenses, the route selection should meet the following constraints:

4.2.2. Time Constraints

Time constraints in travel planning are as follows:

The travel time TT_i is included in the play time of the attraction p_i and the halfway time to the next attraction p_j. The TT_i cannot exceed the travel time limit set by the tourist himself.

Based on the actual situation, it is considered that when each tourist visits the scenic spot, it is affected by the number of tourists in the scenic spot, the degree of road congestion, the walking speed of tourists, and the waiting time. In order to improve the utilization of the trip for the total time constraint, the generated route should satisfy the following constraints:

A trip with a smaller value of time constraint Time (R) is more in line with user expectations. The more in line with user expectations, the higher the quality of the planned tourist routes. To sum up, considering the cost and time constraints, the quality of tourist route planning is evaluated by the value (R) of tourists’ itinerary experience:

In equation (16), the values of the two parameters and satisfy . The specific settings change dynamically according to the tourist flow of the scenic spots and the needs of tourists. For example, during holidays, tourists often hope to enjoy a better travel experience in a limited time, so the value of can be set larger. In the case that users do not have specific requirements, this study will use a 1 : 1 ratio to allocate time and cost and recommend the results to users.

4.3. Realization of Travel Route Selection

Hangzhou, the capital of Zhejiang Province, is a key scenic tourist city and a famous historical and cultural city in China. Hangzhou has a long history. It is not only the birthplace of “Liangzhu Culture” but also one of the seven ancient capitals of my country. It is known as “the land of fish and rice, the house of silk, the tourist attraction, the state of culture, the paradise on Earth” and so on. This study uses the top ten scenic spots in Hangzhou as the research data and tries to study the optimal travel route within the specified time and within the specified cost. The details of each attraction are listed in Table 1.

Table 1 

Basic information of scenic spots.

According to the latitude and longitude of each scenic spot, the distance between the scenic spots is calculated, and the specific calculation formula is as follows:where the parameter R is the radius of the Earth, and the value is 6371 km.

The optimal route planning is often multiobjective constraints, and the constraints considered in this study include the time spent playing and the cost. Through statistics, the play time and cost details of each scenic spot are listed in Table 2.

Table 2 

Attractions time and cost.

Hangzhou is a prosperous city with a large population. Although the traffic is developed, there are too many cars and people, resulting in serious traffic jams. In order to facilitate the calculation, this study does not consider traffic jams, bad weather, and other conditions. Assuming that the distances to and from each scenic spot are the same, the details of the time to and from each scenic spot are listed in Table 3.

Table 3 

Details of time spent between various attractions.

Aiming at the above data, the IGA algorithm is used for mathematical modeling to obtain the optimal tourist route under the constraints of multiple objectives. Among them, the multiobjective constraints mainly include time constraints and cost constraints. The weight between the two constraints is set to 6 : 4. The optimal path is 1-6-4-2-7-9-3-10-5-8. The distance is 381.46 km.

5. Conclusion

The booming tourism industry has resulted in a substantial increase in the number of attractions and tourists. In many attractions, how tourists arrange personalized tourist routes has become very important. The traditional way of tourism requires tourists to make travel strategies in advance. By collecting a large amount of tourist information on various tourism websites, tourists decide the attractions to visit, the order of visiting the attractions, and the mode of transportation. Obviously, this is a very tedious thing. In order to facilitate tourists to rationally plan tourist routes, this study proposes to use the IGA algorithm for optimal tourist route planning. Traditional GAs are prone to problems such as premature convergence and poor local search ability when planning travel routes. In order to solve the above problems, this study proposes the IGA algorithm. IGA proposes three improvements to traditional GA. One is to introduce the ACA to initialize the parameters in the GA. This algorithm is introduced to alleviate the over-reliance of GA on initializing the population and the poor adaptability of individual populations. Second, because the crossover probability parameter in GA has a great influence on the final solution, this study introduces an adaptive strategy to adjust the crossover probability to improve the fitness of the population. Third, considering the weak local search ability of GA and the problem of premature convergence, this study introduces the 2-opt optimization algorithm to improve the quality of the solution. The experimental analysis results verify the effectiveness of the proposed method. However, this study can be further optimized, for example, the goal of introducing the popularity of attractions and shortening the route optimization time.

Data Availability

The labeled data set used to support the findings of this study is available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was supported by the Project of Hebei Social Science Foundation: “Empirical Study on Micro Mechanism of Ecological Civilization Construction of Tourism in Hebei Province” (no. HB17YJ039).