A multi-index optimization space crowdsourcing task allocation method based on a genetic algorithm

Abstract

The application discloses a multi-index optimization space crowd sourcing task allocation method based on a genetic algorithm, and steps of the method comprise the following steps: 1) obtaining arrival time and position information of a task and arrival time and position information of a worker; 2) constructing a task allocation model for a space crowd sourcing platform; 3) realizing and improving application of the genetic algorithm in the space crowd sourcing, and obtaining an optimal task allocation scheme. The application can find the optimal task allocation scheme of task allocation through continuous iteration, so that the efficiency and utility of task allocation of the space crowd sourcing platform can be improved.

Description

Technical Field

[0001] This invention belongs to the field of spatial crowdsourcing task allocation. Specifically, it is a multi-index optimization method for spatial crowdsourcing task allocation based on genetic algorithms. Background Technology

[0002] With the development of smart devices and high-speed wireless networks, spatial crowdsourcing, which assigns location-based tasks to workers, has received widespread attention in recent years. Specifically, workers need to physically move to a designated location to complete tasks posted by task requesters. However, some complex tasks not only require specific skills but also necessitate ensuring workers are within a certain distance to guarantee fair compensation.

[0003] Most existing task allocation schemes in space crowdsourcing assume that all tasks are simple and can be easily completed by one person, such as delivering packages, taking photos, and reporting on trending topics. However, a single worker cannot complete some complex tasks, such as preparing for a party or decorating a house. For example, decorating a house requires workers to design blueprints, paint walls, lay tiles, move furniture, and install plumbing and electrical systems. Therefore, crowdsourcing platforms need to assign multiple workers with different skills to meet the needs of various tasks.

[0004] The spatial crowdsourcing task allocation problem aims to assign suitable workers to each task arriving at the crowdsourcing platform within a certain period. Most existing task allocation schemes optimize for a single metric, such as maximizing worker rewards, minimizing worker travel distance, maximizing the number of tasks assigned, and minimizing worker idle time. However, there is relatively little research on task allocation that comprehensively considers multiple metrics, resulting in task allocation schemes that perform well in one aspect but poorly in others. Summary of the Invention

[0005] The present invention aims to address the shortcomings of the existing technology by proposing a multi-index optimization method for spatial crowdsourcing task allocation based on genetic algorithms. This method seeks to find the optimal task allocation scheme, thereby improving the efficiency and effectiveness of task allocation on spatial crowdsourcing platforms.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0007] The present invention discloses a multi-index optimization method for spatial crowdsourcing task allocation based on genetic algorithms, characterized in that it is applied to a spatial crowdsourcing platform that allocates tasks to users. individual worker Assigned to several arriving Task In the scenario, among which, Indicates the first One worker, and , Indicates the first individual worker Geographical location Indicates the first individual worker Arrival time, Indicates the first individual worker Travel distance, Indicates the first individual worker The skills they possess , Represents a set of skills, and , Indicates the first One skill, and The corresponding reward is recorded as , This represents the total number of skills. Indicates the first individual worker The rewards received Indicates the first One task, and , Indicates the first Task Geographical location Indicates the first Task Arrival time, Indicates the first Task Required skills list ; The total number of workers, The total number of tasks. , The multi-index optimization spatial crowdsourcing task allocation method is carried out according to the following steps:

[0008] Step 1: Construct a multi-index optimization model for spatial crowdsourcing task allocation:

[0009] Step 1.1: Construct the objective function of the multi-index optimization spatial crowdsourcing task allocation model using equation (1). :

[0010] (1)

[0011] In equation (1), This indicates the number of workers completing the task in a single task allocation scheme. Represents the i-th worker The rewards for completing the task This represents the maximum reward value, and satisfies the following conditions: , Represents the i-th worker The distance traveled to complete the task. This represents the maximum acceptable travel distance for a worker in a single task assignment scheme; This represents the success rate of task allocation across all tasks in a single task allocation scheme. ,in This represents the number of tasks successfully assigned in a single task allocation scheme. Total number of tasks; There are 3 weighting coefficients, and , ;

[0012] Step 1.2: Construct the constraints of the multi-index optimization spatial crowdsourcing task allocation model using equations (2)-(4):

[0013] (2)

[0014] (3)

[0015] (4)

[0016] In equation (2), Represents the distance function;

[0017] Step 2: Use a genetic algorithm to solve the multi-index optimization spatial crowdsourcing task allocation model to obtain the optimal task allocation scheme;

[0018] Step 2.1: Define the current iteration round as... and initialize The maximum number of iterations is The deviation threshold of the fitness of the best individual in two adjacent rounds is The population size is ;

[0019] From including all arrivals A task allocation scheme is composed of a task, a set of workers who have been assigned tasks, and a set of idle workers waiting to be assigned tasks. This scheme is then used as an individual to randomly initialize the population of generation G. ,in, Indicates the Gth generation population The Middle Individual;

[0020] Step 2.2: Calculate the first step using equation (1). Generation population The Middle individual fitness Thus, the first Generation population The fitness of each individual is calculated, and the individual with the highest fitness is selected. The corresponding individual is denoted as the first. The optimal individual ;

[0021] Step 2.3, according to Calculate the first individual The probability of being selected And based on the probability of being selected, a roulette wheel selection method is used to select... conduct The next choice, each time from After selecting two individuals for crossover and mutation, two offspring individuals are generated, thus obtaining the first generation. Offspring population ;

[0022] Step 2.4: Calculate using equation (1) The fitness of each offspring individual, and compared with The fitness of each individual is sorted in descending order, and the top... The individual composition corresponding to each fitness level is as follows: Generation population and from Select the maximum fitness The corresponding individual is denoted as the first. The optimal individual ;

[0023] Step 2.5, if or iteration rounds If so, assign G+1 to G and return to step 2.3 for sequential execution; otherwise, ... Output the optimal task allocation scheme.

[0024] The multi-index optimization spatial crowdsourcing task allocation method based on genetic algorithm described in this invention is also characterized in that the crossover operation in step 2.3 is performed as follows:

[0025] Step 2.3.1, in Randomly select the xth individual and the yth individual And randomly select the z-th task As an intersection, thus The zth task The set of assigned workers and The zth task The set of assigned workers exchange;

[0026] Step 2.3.2: After the swap, the xth individual... In Add the xth individual to the set of workers with available tasks awaiting assignment. The workers who have been swapped and whose tasks have been repeatedly assigned, as well as the idle workers waiting to be assigned tasks, are concentrated from the xth individual. Workers who did not intersect are deleted from the corresponding set, thus obtaining the xth intersecting individual. ;

[0027] The y-th individual In Add the y-th individual to the set of idle workers awaiting assignment. The workers, after being swapped, consist of workers who have already been assigned tasks and idle workers awaiting assignment, from... Workers who did not intersect are deleted from the corresponding group, thus obtaining the y-th intersecting individual. ;

[0028] Step 2.3.3: Traverse the x-th intersection individual If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the x-th updated crossover individual. ;

[0029] Traversal If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the y-th updated crossover individual. .

[0030] Furthermore, the mutation operation in step 2.3 is performed as follows:

[0031] from Randomly select the zz-th task and its corresponding worker set And in accordance with the constraints, from Idle workers awaiting task assignment are Workers will be reassigned. If no worker meeting the constraints is found, then... The worker set is assigned an empty set, and the mutation operation is performed before... Workers in the field joined The idle workers awaiting assignment are gathered together to obtain a mutated offspring individual. .

[0032] The present invention provides an electronic device, comprising a memory and a processor, wherein the memory is used to store a program that supports the processor in executing the multi-index optimization spatial crowdsourcing task allocation method, and the processor is configured to execute the program stored in the memory.

[0033] The present invention discloses a computer-readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, performs the steps of the multi-index optimization spatial crowdsourcing task allocation method.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0035] 1. This invention matches the skills possessed by workers with the numerous skill requirements of the task and adds corresponding constraints for task allocation, thereby flexibly responding to complex single-task multi-worker scenarios and better meeting the needs of both tasks and workers.

[0036] 2. In the spatial crowdsourcing task allocation stage, this invention introduces an objective function that comprehensively considers multiple indicators, namely maximizing worker rewards, minimizing worker travel distance, and maximizing task allocation success rate. This objective function is used as the fitness calculation function for individuals, thereby increasing worker rewards, reducing worker travel distance, and improving task allocation success rate, making reasonable allocations, and outputting the optimal task allocation scheme. Attached Figure Description

[0037] Figure 1 This is a flowchart of the genetic algorithm of the present invention;

[0038] Figure 2 This is a conceptual diagram of the population and individuals in the genetic algorithm of this invention;

[0039] Figure 3 This is a schematic diagram of the crossover operation in the genetic algorithm of this invention;

[0040] Figure 4 This is a schematic diagram of the mutation operation in the genetic algorithm of this invention;

[0041] Figure 5 This is a flowchart of the method of the present invention. Detailed Implementation

[0042] In this embodiment, a multi-index optimization method for spatial crowdsourcing task allocation based on genetic algorithms is a task allocation method for spatial crowdsourcing. It combines real-life scenarios, considering the arrival time and location information of tasks and workers, to find the optimal task allocation scheme that maximizes worker rewards and task allocation success rate while minimizing worker travel distance. Specifically, it is applied to scenarios where a spatial crowdsourcing platform allocates several workers to several arriving tasks, such as... Figure 5 As shown, proceed as follows:

[0043] Step 1: The space crowdsourcing platform will individual worker Assigned to several arriving Task In the scenario, among which, Indicates the first One worker, and , Indicates the first individual worker Geographical location Indicates the first individual worker Arrival time, Indicates the first individual worker Travel distance, Indicates the first individual worker The skills they possess , Represents a set of skills, and , Indicates the first One skill, and The corresponding reward is recorded as , This represents the total number of skills. Indicates the first individual worker The rewards received Indicates the first One task, and , Indicates the first Task Geographical location Indicates the first Task Arrival time, Indicates the first Task Required skills list ; The total number of workers, The total number of tasks. , .

[0044] Step 2: Construct a multi-index optimization model for spatial crowdsourcing task allocation:

[0045] Step 2.1: Construct the objective function of the multi-index optimization spatial crowdsourcing task allocation model using equation (1). :

[0046] (1)

[0047] Equation (1) represents the sum of maximizing worker rewards, minimizing worker travel distance, and maximizing task assignment success rate; in Equation (1), This indicates the number of workers completing the task in a single task allocation scheme. Represents the i-th worker The rewards for completing the task This represents the maximum reward value, and satisfies the following conditions: , Represents the i-th worker The distance traveled to complete the task. This represents the maximum acceptable travel distance for a worker in a single task assignment scheme; This represents the success rate of task allocation across all tasks in a single task allocation scheme. ,in This represents the number of tasks successfully assigned in a single task allocation scheme. Total number of tasks; There are 3 weighting coefficients, and , .

[0048] Step 2.2: Construct the constraints of the multi-index optimization spatial crowdsourcing task allocation model using equations (2)-(4):

[0049] (2)

[0050] (3)

[0051] (4)

[0052] In equation (2), Representing a distance function, workers do not accept distance in... Other tasks;

[0053] Equation (3) indicates that only workers who arrive before the task can be assigned tasks;

[0054] Equation (4) indicates that the worker assigned the task must possess the skills required for that task.

[0055] Step 3: Use a genetic algorithm to solve the multi-index optimization spatial crowdsourcing task allocation model to obtain the optimal task allocation scheme, such as... Figure 1 As shown, the genetic algorithm will proceed in the following steps:

[0056] Step 3.1: Define the current iteration round as... and initialize The maximum number of iterations is The deviation threshold of the fitness of the best individual in two adjacent rounds is The population size is ;

[0057] From including all arrivals A task allocation scheme is composed of a task, a set of workers who have been assigned tasks, and a set of idle workers waiting to be assigned tasks. This scheme is then used as an individual to randomly initialize the population of generation G. ,in, Indicates the Gth generation population The Middle Individual; Conceptual diagram of population and individual, as shown in the figure. Figure 2 As shown, individuals 1, 2, etc., form a population, and individual 1 represents a task allocation scheme, task... If no workers are assigned, then an empty set is assigned. ,Task The assigned worker sets are respectively The set of idle workers awaiting assignment is .

[0058] Step 3.2: Calculate the first step using equation (1). Generation population The Middle individual fitness Thus, the first Generation population The fitness of each individual is calculated, and the individual with the highest fitness is selected. The corresponding individual is denoted as the first. The optimal individual ;

[0059] Step 3.3, according to Calculate the first individual The probability of being selected And based on the probability of being selected, a roulette wheel selection method is used to select... conduct The next choice, each time from Select two individuals to perform crossover and mutation operations.

[0060] Step 3.4, Cross Operation:

[0061] Step 3.4.1, in Randomly select the xth individual and the yth individual And randomly select the z-th task As an intersection, thus The zth task The set of assigned workers and The zth task The set of assigned workers exchange;

[0062] Step 3.4.2: After the swap, the xth individual... In Add the xth individual to the set of workers with available tasks awaiting assignment. The workers who have been swapped and whose tasks have been repeatedly assigned, as well as the idle workers waiting to be assigned tasks, are concentrated from the xth individual. Workers who did not intersect are deleted from the corresponding set, thus obtaining the xth intersecting individual. ;

[0063] The y-th individual In Add the y-th individual to the set of idle workers awaiting assignment. The workers, after being swapped, consist of workers who have already been assigned tasks and idle workers awaiting assignment, from... Workers who did not intersect are deleted from the corresponding group, thus obtaining the y-th intersecting individual. ;

[0064] Step 3.4.3: Traverse the x-th intersection individual If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the x-th updated crossover individual. ;

[0065] Traversal If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the y-th updated crossover individual. .

[0066] Assume the task of parent generation 1 The assigned worker sets are respectively The set of idle workers awaiting assignment is The worker sets for task assignment in parent generation 2 are respectively The set of idle workers awaiting assignment is Specifically, a task is randomly selected from all arriving tasks, based on a uniform distribution. For example, task... Task in Parent Generation 1 workers Tasks assigned to parent generation 2 Parent Generation 2 Task workers Tasks assigned to parent generation 1 ;

[0067] After the worker sets of the two parent tasks are swapped, parent 1 produces child 1, and parent 2 produces child 2. If there are some duplicate workers, they are integrated, prioritizing the worker sets that have been swapped. For example, in child 2, the tasks... Workers exchanged Compared to the original sub-generation 2 mission workers There are repetitive workers , and idle workers awaiting assignment There are repetitive workers In this case, the worker set that was swapped in will be retained first, and the tasks will be deleted. workers workers Remove workers from the group of idle workers awaiting assignment. ;

[0068] Then return the workers who were swapped out to the pool of workers awaiting assignment. For example, in parent generation 2 task... Workers exchanged Return it to the set of idle workers to be assigned in child generation 2. In the middle, idle workers awaiting assignment become... ;

[0069] The removal of duplicate workers resulted in some assigned workers being missing from tasks that originally had worker sets, such as tasks in child generation 2. workers The task was deleted because it was a duplicate of the worker set that was swapped, resulting in... The worker set becomes an empty set Search among the available workers awaiting assignment for suitable tasks. If a worker with constraints can be found, then it is considered a task. Assign workers; if no suitable workers can be found, then the task... Workers' Collection of Employment Space This indicates that the task assignment failed.

[0070] Taking parent generation 1 and parent generation 2 as an example, the process of generating child generation 1 and child generation 2 through crossover is as follows: Figure 3 As shown.

[0071] Step 3.5, Mutation Operation:

[0072] from Randomly select the zz-th task and its corresponding worker set And in accordance with the constraints, from Idle workers awaiting task assignment are Workers will be reassigned. If no worker meeting the constraints is found, then... The worker set is assigned an empty set, and the mutation operation is performed before... Workers in the field joined The idle workers awaiting assignment are gathered together to obtain a mutated offspring individual. .

[0073] Assuming the parent's task The assigned worker sets are respectively Idle workers awaiting assignment are grouped as Specifically, a task is randomly selected from all arriving tasks, based on a uniform distribution. For example, task... Clear tasks already assigned workers Find suitable tasks from among the available workers waiting to be assigned. If a set of workers that meets the constraints is found, then... This will gather new workers. Assigned to task Afterwards, the original workers were gathered Return the workers to the set of available workers awaiting assignment; if no suitable set of workers is found, simply return the original set of workers. Return the idle workers to the designated area for assignment; task The workers set empty sets, and consider the task Allocation failed;

[0074] The process of generating offspring through mutation operations is as follows: Figure 4 As shown.

[0075] Step 3.6: Calculate using equation (1) The fitness of each offspring individual, and compared with The fitness of each individual is sorted in descending order, and the top... The individual composition corresponding to each fitness level is as follows: Generation population and from Select the maximum fitness The corresponding individual is denoted as the first. The optimal individual ;

[0076] Step 3.7, if or iteration rounds If so, assign G+1 to G and return to step 3.2 for sequential execution; otherwise, ... Output the optimal task allocation scheme.

[0077] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the above-described method, and the processor is configured to execute the program stored in the memory.

[0078] In this embodiment, a computer-readable storage medium stores a computer program, which is executed by a processor to perform the steps of the above method.

Claims

1. A multi-index optimization spatial crowdsourcing task allocation method based on genetic algorithm, characterized in that, It is applied to a space crowdsourcing platform individual worker Assigned to several arriving Task In the scenario, among which, Indicates the first One worker, and , Indicates the first individual worker Geographical location Indicates the first individual worker Arrival time, Indicates the first individual worker Travel distance, Indicates the first individual worker The skills they possess , Represents a set of skills, and , Indicates the first One skill, and The corresponding reward is recorded as , This represents the total number of skills. Indicates the first individual worker The rewards received Indicates the first One task, and , Indicates the first Task Geographical location Indicates the first Task Arrival time, Indicates the first Task Required skills list ; The total number of workers, The total number of tasks. , The multi-index optimization spatial crowdsourcing task allocation method is carried out according to the following steps: Step 1: Construct a multi-index optimization model for spatial crowdsourcing task allocation: Step 1.1: Construct the objective function of the multi-index optimization spatial crowdsourcing task allocation model using equation (1). : (1) In equation (1), This indicates the number of workers completing the task in a single task allocation scheme. Represents the i-th worker The rewards for completing the task This represents the maximum reward value, and satisfies the following conditions: , Represents the i-th worker The distance traveled to complete the task. This represents the maximum acceptable travel distance for a worker in a single task assignment scheme; This represents the success rate of task allocation across all tasks in a single task allocation scheme. ,in This represents the number of tasks successfully assigned in a single task allocation scheme. Total number of tasks; There are 3 weighting coefficients, and , ; Step 1.2: Construct the constraints of the multi-index optimization spatial crowdsourcing task allocation model using equations (2)-(4): (2) (3) (4) In equation (2), Represents the distance function; Step 2: Use a genetic algorithm to solve the multi-index optimization spatial crowdsourcing task allocation model to obtain the optimal task allocation scheme; Step 2.1: Define the current iteration round as... and initialize The maximum number of iterations is The deviation threshold of the fitness of the best individual in two adjacent rounds is The population size is ; From including all arrivals A task allocation scheme is composed of a task, a set of workers who have been assigned tasks, and a set of idle workers waiting to be assigned tasks. This scheme is then used as an individual to randomly initialize the population of generation G. ,in, Represents the Gth generation population The Middle Individual; Step 2.2: Calculate the first step using equation (1). Generation population The Middle individual fitness Thus, the first Generation population The fitness of each individual is calculated, and the individual with the highest fitness is selected. The corresponding individual is denoted as the first. The optimal individual ; Step 2.3, according to Calculate the first individual The probability of being selected And based on the probability of being selected, a roulette wheel selection method is used to select... conduct The next choice, each time from After selecting two individuals for crossover and mutation, two offspring individuals are generated, thus obtaining the first generation. Offspring population ; Step 2.4: Calculate using equation (1) The fitness of each offspring individual, and compared with The fitness of each individual is sorted in descending order, and the top... The individual composition corresponding to each fitness level is as follows: Generation population and from Select the maximum fitness The corresponding individual is denoted as the first. The optimal individual ; Step 2.5, if or iteration rounds If so, assign G+1 to G and return to step 2.3 for sequential execution; otherwise, ... Output the optimal task allocation scheme.

2. The multi-index optimization spatial crowdsourcing task allocation method based on genetic algorithm according to claim 1, characterized in that, The crossover operation in step 2.3 is performed as follows: Step 2.3.1, in Randomly select the xth individual and the yth individual And randomly select the z-th task As an intersection, thus The zth task The set of assigned workers and The zth task The set of assigned workers exchange; Step 2.3.2: After the swap, the xth individual... In Add the xth individual to the set of workers with available tasks awaiting assignment. The workers who have been swapped and whose tasks have been repeatedly assigned, as well as the idle workers waiting to be assigned tasks, are concentrated from the xth individual. Workers who did not intersect are deleted from the corresponding set, thus obtaining the xth intersecting individual. ; The y-th individual In Add the y-th individual to the set of idle workers awaiting assignment. The workers, after being swapped, consist of workers who have already been assigned tasks and idle workers awaiting assignment, from... Workers who did not intersect are deleted from the corresponding group, thus obtaining the y-th intersecting individual. ; Step 2.3.3: Traverse the x-th intersection individual If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the x-th updated crossover individual. ; Traversal If any tasks in the task list are unassigned, then those unassigned tasks will be assigned to other tasks according to the constraints. For idle workers awaiting task assignment, if no worker matching the constraints is found, the worker set for the corresponding task is set to empty, thus obtaining the y-th updated crossover individual. .

3. The multi-index optimization spatial crowdsourcing task allocation method based on genetic algorithm according to claim 1, characterized in that, The mutation operation in step 2.3 is performed as follows: from Randomly select the zz-th task and its corresponding worker set And in accordance with the constraints, from Idle workers awaiting task assignment are Workers will be reassigned. If no worker meeting the constraints is found, then... The worker set is assigned an empty set, and the mutation operation is performed before... Workers in the field joined The idle workers awaiting assignment are gathered together to obtain a mutated offspring individual. .

4. An electronic device, comprising a memory and a processor, characterized in that, The memory is used to store programs that support the processor in executing the multi-index optimization spatial crowdsourcing task allocation method of any one of claims 1-3, and the processor is configured to execute the programs stored in the memory.

5. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is run by the processor, it executes the steps of the multi-index optimization spatial crowdsourcing task allocation method according to any one of claims 1-3.

Images