Mass test task allocation method based on task dynamic priority

Abstract

According to the crowd test task allocation method based on the task dynamic priority, the crowd test task allocation model and the crowd test task allocation algorithm based on the task dynamic priority are provided in consideration of the importance difference between the tasks, the overall task completion quality is greatly improved while high-priority tasks are taken into consideration, and the task allocation efficiency is improved. And a new thought is provided for the public test platform to realize efficient task allocation. The method is mainly characterized in that 1) based on the EDF thought, an initial priority sequence of tasks is constructed according to budget, number of people and time requirements of the tasks, and relative weights of different factors are determined by using an analytic hierarchy process (AHP); and 2) with maximization of the overall matching degree S as a target, establishing a public test task allocation model based on the dynamic priority. And 3) according to a greedy thought, finding the task with the highest priority in the task allocation priority sequence during each allocation, and then allocating a tester with the highest matching degree to the task according to the matching degree matrix. And 4) according to the testee characteristics, updating the testee number demand and the priority of the task, and updating the priority sequence. And performing cyclic distribution of the personnel according to a priority rule until a distribution termination condition is reached.

Description

Technical Field

[0001] The present invention belongs to the field of software crowdsourcing testing and involves Web application technology, and in particular, to the field of task allocation for functional testing of a crowdsourcing testing platform. Taking into account the different importance of tasks, a crowd testing task allocation model based on dynamic task priority and its allocation algorithm are proposed. While taking care of high-priority tasks, the overall task completion quality is greatly improved, providing a new idea for the crowd testing platform to achieve efficient task allocation. Background Art

[0002] Crowdsourcing testing is a promising new software testing method that uses crowdsourcing ideas and adopts an open innovation process. In other words, crowdsourcing testing is the application of crowdsourcing mechanism in the software testing industry. For the current popular distributed software, compared with traditional software, its large scale and complex system further increase the difficulty of testing, which also means that more professionals need to be hired. Crowdsourcing testing greatly shortens the testing cycle and significantly reduces the testing cost by recruiting a large number of people online in a short period of time to participate in the testing. At the same time, it largely solves the problems faced in the testing process, such as insufficient test diversity, less feedback, and unrealistic application scenarios. However, it also poses new challenges to the management of crowdsourcing testing.

[0003] In task scheduling, it is an important means to optimize the task allocation process by calculating the task priority. Commonly used methods are the Least Slack Time First (LSF) method and the Earliest Deadline First (EDF) method and their improved algorithms. The former determines the task scheduling priority based on the minimum idle value of the remaining execution time, and the latter uses the upper limit of the deadline to determine the real-time task scheduling priority. In the field of crowdsourcing testing, the determination of task priority may include factors such as amount, number of people, and time. Crowdsourcing tasks with higher priorities often mean higher value or more urgent needs. Therefore, from this perspective, the reasonable calculation of crowdsourcing task priorities can help distinguish the differences in importance between tasks. At the same time, giving priority to the completion quality of high-priority tasks in the process of task allocation will attract more high-quality users to register, which is more conducive to the long-term development of the platform. Summary of the invention

[0004] The problem to be solved by the present invention is to optimize the crowdsourcing test task allocation process based on task dynamic priority technology and improve the test quality.

[0005] Definition 1: Task Value

[0006] Refers to the possible benefits that the completion of a task can bring to the crowd-testing platform. This article measures the value of a task by the expected total budget of the task and the expected number of testers. When the expected number of testers for a task is the same, the higher the expected total budget of the task, the higher the value of the task. i , whose task value is u value for:

[0007]

[0008] where u money is the expected total budget for the task, u cnt The expected number of testers for the task.

[0009] Definition 2: Task Urgency

[0010] Describes the time urgency of task assignment. This article measures the value of a task by the expected total budget and the expected number of testers. When the expected number of testers is the same, the shorter the expected completion time, the higher the urgency of the task. i , the task urgency u urgency for:

[0011]

[0012] where u time is the expected completion time of the task, u cnt The expected number of testers for the task.

[0013] Definition 3: Initial priority of tasks

[0014] Refers to the priority of a task before it is assigned. The priority of a task is determined by its value u value and the urgency of the task u urrgency Comprehensive decision. For the test task u i , its initial task priority P i for:

[0015]

[0016] Among them, γ and 1-γ represent the weight values ​​of task value and task urgency respectively, and 0<γ<1. The analytic hierarchy process (AHP) is used for evaluation.

[0017] Define 4 task dynamic priorities

[0018] Considering that each tester has different characteristics and has different impacts on task completion, each tester will have different impacts on task priority during the actual assignment process. This paper introduces the priority factor μ, combined with the average matching degree of the tester to all tasks, to achieve dynamic changes in task priority.i Assigned to tester w j When task u i The priority of the task is updated to obtain the dynamic priority P i ′:

[0019]

[0020] Where P i is the initial priority of the task, For tester w j For the average matching value of n tasks to be assigned. cnt is the expected number of testers for the task, and μ is the priority factor and μ≤1.

[0021] Definition 5: Crowd-testing task allocation model based on dynamic priority

[0022] In the constraint condition, the priority of the task in a single allocation must be greater than or equal to the highest priority P′ of all tasks at this time. max , ensuring that each allocation is allocated to the task with the highest priority at the moment. That is, given the crowd-testing task set u, the candidate tester set w, and the task-tester matching matrix S, given the initial priority of the task, from any possible allocation scheme, under the premise of ensuring the quality of the high-priority task, seek an allocation scheme that maximizes the overall matching degree. Decision variable X ij Defined as whether to put the i-th test task u i Assigned to the jth tester w j When X ij =1, it means that the i-th test task u i Assigned to the jth tester w j , when X ij =0 means no allocation.

[0023]

[0024] The technical solution of the present invention is:

[0025] 1) Based on the EDF concept, the initial priority sequence of tasks is constructed according to the budget, number of people and time requirements of the tasks themselves, and the relative weights of different factors are determined using the analytic hierarchy process (AHP).

[0026] 2) With the goal of maximizing the overall matching degree S, the following crowd testing task allocation model based on dynamic priority is established.

[0027] 3) According to the greedy idea, each time the task is assigned, the highest priority task is found in the task assignment priority sequence, and then the tester with the highest matching degree is assigned to it according to the matching degree matrix.

[0028] 4) Update the number of testers and the priority of the task according to the tester characteristics and update the priority sequence. Perform cyclic allocation of personnel according to the priority rules until the termination allocation condition is reached.

[0029] The beneficial effect of the present invention is that the traditional task priority allocation method often only considers the matching degree of the task-executor to give priority to the high-priority tasks, which will lead to the emergence of resource absorption effect, that is, high-priority tasks are all performed by high-capability personnel, and low-priority tasks are all performed by low-capability personnel, which will lead to problems such as low completion quality of some tasks and uneven distribution of testers. In response to this phenomenon, the present invention balances resource allocation by realizing dynamic changes in task priorities during the allocation process, so that low-priority tasks can also obtain a certain number of high-quality testers without affecting the overall completion quality, further effectively ensuring the quality of crowdsourcing testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 Business process diagram

[0031] Figure 2 EDF workflow diagram

[0032] Figure 3 AHP workflow diagram Specific implementation methods

[0034] In the process of task allocation, this method calculates the initial priority sequence of the task based on the budget, number of people and time requirements of the task itself. Then, according to the greedy idea, personnel are allocated one by one. Each time the task is allocated, the highest priority task is found in the task allocation priority sequence, and then the tester with the highest matching degree is allocated to it according to the matching matrix. Then, according to the characteristics of the tester, the number of people required and the priority of this task are updated, and the priority sequence is updated. According to the priority rules, personnel are allocated cyclically until the termination condition is reached. The business process of this method is as follows: Figure 1 shown.

[0035] After obtaining the matching matrix of all tasks and testers according to the matching algorithm, we use it as one of the inputs and first calculate the value of each task u according to the expected total budget and the expected number of testers. value Then calculate the urgency of the task u according to the expected completion time and the expected number of testers urgency , and then calculate the initial priority P of each task based on the task value and task urgency i Then, the cycle starts, finds the task with the highest priority, and selects the tester with the highest matching degree from the candidate list. j Before each assignment, the task needs to be judged. If the task has reached the number of people required, cnt, then its priority P i ′ is set to minus one, which means that no further allocation is required and the next task is postponed. If all the candidates in the candidate list have been allocated, the allocation process ends. If the allocation is successful, the remaining number of people required for the task is reduced by one, and the priority of the task is updated according to the average matching degree of the testers allocated this time and inserted into the task priority sequence for the next allocation. After each personnel allocation, the termination condition is judged according to the allocation matrix, that is, all tasks meet their number requirements or all testers have been allocated, and finally a 0-1 matching matrix is ​​output.

[0036] The flowchart of the Earliest Deadline First (EDF) strategy is as follows: Figure 2 , which specifically includes the following steps:

[0037] Step 1: Task arrival: When a new task arrives, it is added to the task queue.

[0038] Step 2: Task sorting: Sort the task queue according to the task deadline, with the task with the earliest deadline at the front.

[0039] Step 3: Task scheduling: Select the task with the earliest deadline for execution.

[0040] Step 4: Task Completion: When a task is completed, it is removed from the task queue.

[0041] Step 5: Repeat steps 1-4: Keep repeating the above steps until all tasks are completed.

[0042] The flowchart of the analytic hierarchy process (AHP) is as follows: Figure 3 , which specifically includes the following steps:

[0043] Step 1: Establish a hierarchy: including the goal layer: the ultimate goal of the decision, the criteria layer: the criteria or standards that affect the goal,

[0044] Plan layer: optional decision-making plans.

[0045] Step 2: Construct a pairwise comparison matrix: Compare the elements of each layer pairwise to construct a pairwise comparison matrix.

[0046] Step 3: Calculate the weight vector: Perform eigenvalue decomposition on the pairwise comparison matrix and calculate the weight vector.

[0047] Step 4: Consistency test: Calculate the consistency ratio (CR) to test the consistency of the pairwise comparison matrix.

[0048] Step 5: Comprehensive evaluation: Calculate the comprehensive score of each plan and select the best plan.

[0049] In summary, this method constructs a priority-based crowd testing task allocation model by introducing the concept of task priority, and realizes the effective allocation of crowd testing tasks based on dynamic task priorities.

Claims

1. A crowd testing task allocation method based on task dynamic priority, characterized by: (1) Based on the EDF concept, the initial priority sequence of the task is constructed according to the budget, number of people and time requirements of the task itself, and the relative weights of different factors are determined using the analytic hierarchy process (AHP). (2) With the goal of maximizing the overall matching degree S, the following crowd-testing task allocation model based on dynamic priority is established. (3) According to the greedy concept, each time the task is allocated, the highest priority task is found in the task allocation priority sequence, and then the tester with the highest matching degree is assigned to it according to the matching degree matrix. (4) According to the characteristics of the tester, the number of testers required and the priority of the task are updated, and the priority sequence is updated. The personnel are allocated cyclically according to the priority rules until the termination condition is reached.

2. A crowd-testing task allocation method based on task dynamic priority according to claim 1, characterized in that Identify the factors that influence task priority. In the field of crowdsourcing testing, task priority may be determined by factors such as amount, number of people, and time. Crowdsourcing testing tasks with higher priorities often mean higher value or more urgent needs. From this perspective, reasonable calculation of crowdsourcing task priorities helps distinguish the importance of tasks.

3. A crowd-testing task allocation method based on task dynamic priority according to claim 1, characterized in that In the task allocation model, the importance of the task is taken into consideration, the priority of all tasks is calculated, and the task priority is affected by the allocation of tasks through relevant restrictions, so as to ensure that high-priority tasks have priority to obtain higher expected completion quality while making the overall task completion quality higher.

4. A method for allocating crowd-testing tasks based on dynamic task priority according to claim 1, characterized in that After obtaining the matching matrix of all tasks and testers according to the matching algorithm, we use it as one of the inputs and first calculate the value of each task u according to the expected total budget and the expected number of testers. value Then calculate the urgency of the task u according to the expected completion time and the expected number of testers urgency .

5. A crowd-testing task allocation method based on task dynamic priority according to claim 1, characterized in that Calculate the initial priority P of each task based on task value and task urgency i Then, the cycle starts, finds the task with the highest priority, and selects the tester with the highest matching degree from the candidate list. j Make an allocation.

6. A crowd-testing task allocation method based on task dynamic priority according to claim 1, characterized in that Before each assignment, the task needs to be judged. If the task has reached the number of people required, cnt , then its priority P i ′ is set to minus one, which means that no further allocation is required and the next task is postponed. If all the candidates in the candidate list have been allocated, the allocation process ends. If the allocation is successful, the remaining number of people required for the task is reduced by one. At the same time, the priority of the task is updated according to the average matching degree of the testers allocated this time and inserted into the task priority sequence for the next allocation. After each personnel allocation, the termination condition is judged according to the allocation matrix, that is, all tasks meet their number requirements or all testers have been allocated, and finally a 0-1 matching matrix is ​​output.

Images