CN107317872B - Scheduling method of multi-type tasks in space crowdsourcing - Google Patents
Scheduling method of multi-type tasks in space crowdsourcing Download PDFInfo
- Publication number
- CN107317872B CN107317872B CN201710595827.5A CN201710595827A CN107317872B CN 107317872 B CN107317872 B CN 107317872B CN 201710595827 A CN201710595827 A CN 201710595827A CN 107317872 B CN107317872 B CN 107317872B
- Authority
- CN
- China
- Prior art keywords
- task
- scheduling
- tasks
- branch
- candidate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/52—Network services specially adapted for the location of the user terminal
Abstract
The invention discloses a scheduling method of multi-type tasks in space crowdsourcing, which plans a task execution path aiming at tasks distributed to each user, combines professional matching scores of the multi-type tasks, and adopts a scheduling algorithm based on a branch-and-bound thought, but when the task scale is large, the scheduling algorithm has a low operation speed. Two approximate scheduling algorithms, the most promising branch heuristic algorithm and the heuristic algorithm based on the width k search, can be adopted, and the balance between the running speed and the accuracy is obtained by adjusting the k value. The method can improve the quality and quantity of task completion, and has advantages in terms of operation speed and accuracy.
Description
Technical Field
The invention relates to a scheduling method of multi-type tasks in spatial crowdsourcing, and belongs to the technical field of distributed computing.
Background
Crowdsourcing refers to a "distributed problem solving model" that groups perform work tasks performed by full-time employees in the past by outsourcing them to unspecified solution providers in a voluntary fashion through a public Web platform. The early question-and-answer platform Yahoo! Answers, wikipedia, hundredth knowledge, etc. are all crowdsourcing platforms. In the later period, large crowdsourcing platforms such as AmazonNuclear disks, CrowdFlower, oDesk and the like appear. With the popularization of smart phones and the improvement of wireless broadband networks, a new development direction of Crowdsourcing, namely Spatial Crowdsourcing (Spatial crowdssourcing), is formed. In crowd-sourcing, a user can perform a task only by performing the task online and submitting the results of the task, while in spatial crowd-sourcing, the user must move to a location specified by the task to perform the task. The space crowdsourcing is widely applied To the fields of O2O (Online To Offline), disaster monitoring, traffic management, public safety, logistics management, social media and the like. In the application of real-time O2O, a special car service platform Uber popular in recent years drops for travel and the like, and a special car user is a task publisher and publishes specific taxi taking information; the special car driver is a crowd-sourced participant and is responsible for delivering the customer to a destination. In the spatial crowdsourcing application aiming at obtaining data, the U.S. Gigwalk company collects prices of supermarket commodities by using users with smart phones through a spatial crowdsourcing platform. Therefore, the application of space crowdsourcing has invisibly merged into our life and plays an increasingly important role in life. Not only brings the revolution of new technology, but also brings huge economic value, and has attracted wide attention of academic and industrial circles.
In the field of spatial crowdsourcing, a core problem is spatial task allocation. Taking the service type O2O application as an example, a guest (spatial crowdsourcing task publisher) publishes required services (spatial crowdsourcing tasks) such as maintenance of appliances, massage, hairdressing, car washing, etc., and a crowdsourcing platform distributes appropriate service personnel (spatial crowdsourcing users) to the guest, and the service personnel provide services to a place designated by the guest. The service personnel can accept multiple service requirements and then plan routes for all services, performing as many services as possible that have been received. The conventional task scheduling method is discussed aiming at single type task scheduling in a static mode, which is contradictory to the timeliness required in an actual spatial crowdsourcing platform. Tasks in space crowdsourcing require users to complete in time, for example, hairdressing services scheduled by clients, service personnel arrive at a later time to influence the evaluation of the users on the services, and the timeliness of a space task scheduling algorithm is required to be high. The invention provides a method for planning a reasonable execution path aiming at the characteristics of multiple types of tasks and the execution sequence problem of the multiple tasks allocated to the user in a server task pushing mode, thereby maximizing the number of completed tasks and reducing the moving cost as much as possible.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention provides a scheduling method of multiple types of tasks in spatial crowdsourcing, which can improve the quality and the quantity of task completion and has advantages in the aspects of operation speed and accuracy.
Defining 1 Spatial Task (Spatial Task) A Spatial Task t is of the form < l, d, ei>, wherein l represents the sameThe location of the task, d denotes the deadline of the task t, eiIndicating the type of task t and the subscripts indicating the different types of tasks, the subscripts of the multi-type tasks, i.e. all tasks, being different.
eiThe type of the task is indicated, and the task is related to maintenance of electric appliances, hairdressing, massage and the like. The user can complete the task only when moving to the task position l before the deadline time d of the task, otherwise, the task is expired and invalid, and the invention ignores the execution time required by the task.
Define 2 users (Worker) A user w is of the form < l, q, D, E >, where l represents the current location of the user, q represents the number of tasks that the user can accept at most, D represents the user area, and E represents the set of professional abilities that the user has.
E denotes a set of professional abilities that a user has, and a user may possess a plurality of skills, E ═ E {1,e2,...,ei}. Such as one user being good at hairdressing, massaging, etc.
Define 3 professional Match (expert Match) A Match is of the form < w, t > -where task t is to be in the area of the user when the user has professional ability to complete task t, i.e. eiE, called professional matching; if it is
It is called non-professional matching.
Defining 4 match scores (Score) A Score value Score (w, t) is defined for each match < w, t > representing the quality of the user's task completed.
Obviously, the task of maintaining the electric appliance is distributed to the users with maintenance capability, the completion quality is higher than that of the general users, and therefore, the professional matching score is higher than the non-professional matching score. It is not assumed herein that the professional matching score is 3 and the non-professional matching score is 1.
Define 5 move Cost (Travel Cost) in terms of Euclidean distance d (w)j,tk) Computing a representation user wjAnd spatial task tkThe movement cost of (2).
Define 6 Arrival Time (Arrival Time) given an online applicationUser w and n tasks S assigned to it, task sequence R ═ t1,t2,...,tr) When for i ≠ j, r ≦ n, ti∈S,ti≠tjTo the task t in the task set SiThe time of (d) is defined as:
wherein c (a, b) represents a moving distance, and for convenience of representing the moving distance of the user and the task deadline, the moving distance is represented by Manhattan distance (Manhattan distance), and the time of entering the system by the user is defaulted to be 0.
Define 8 the Upper Bound (Upper Bound) which represents the maximum scheduling score that can be achieved under this branch, by which the search tree is pruned. The upper bound ub _ R calculation for node R is given by:
where curScore represents the score the user gets in reaching the current node R,
the maximum match score that the table might obtain after node R.
The technical scheme is as follows: a method for scheduling multi-type tasks in space crowdsourcing comprises the steps that under a task pushing mode of a server, research objects are multi-type tasks, each task has characteristics, and a task execution path is planned according to tasks allocated to a user; on the basis of the branch-and-bound scheduling method, two approximate scheduling methods are adopted, namely a most promising branch heuristic algorithm and a heuristic algorithm based on width k search.
In the multi-type task scheduling stage based on path planning, the existing branch-and-bound thought algorithm is improved to adapt to the multi-type task scheduling environment, the characteristics of the multi-type tasks are fully combined, the task scheduling order is reasonably planned, and the algorithm can obtain an accurate execution path solution;
secondly, the operation efficiency is improved on the basis of a branch and bound thought algorithm, and two approximate scheduling algorithms, namely a most promising branch heuristic algorithm and a heuristic algorithm based on width k search, are adopted.
Has the advantages that: compared with the prior art, the scheduling method of the multi-type tasks in the space crowdsourcing improves the operation efficiency of the scheduling algorithm based on the branch and the bound, can improve the quality and the quantity of task completion, and has advantages in the aspects of operation speed and accuracy.
Drawings
FIG. 1 is a diagram of task scheduling according to an embodiment of the present invention;
FIG. 2 is a search tree diagram based on branch-and-bound max score scheduling according to an embodiment of the present invention;
FIG. 3 is a workflow diagram of computing a set of candidate tasks according to an embodiment of the invention;
fig. 4 is a flowchart of a scheduling algorithm based on branch-and-bound idea according to an embodiment of the present invention.
Detailed Description
The present invention is further illustrated by the following examples, which are intended to be purely exemplary and are not intended to limit the scope of the invention, as various equivalent modifications of the invention will occur to those skilled in the art upon reading the present disclosure and fall within the scope of the appended claims.
The method for scheduling the multi-type tasks in the space crowdsourcing comprises the following steps:
the method comprises the following steps: calculating a candidate task set, as shown in fig. 3, the specific implementation steps are as follows:
(1) firstly, a current task sequence, a candidate task set and a task to be expanded are obtained, the candidate set of the task to be expanded is initialized, and a null value is assigned.
(2) Traversing all tasks in the current candidate set, checking whether the tasks in the candidate set can be completed within the deadline, and if the tasks can be completed within the deadline, adding the tasks into the next candidate set; otherwise, the task can not be completed in the effective time, and the task is discarded.
For example, the search tree obtained by the KM algorithm in FIG. 2, task t at the first level5Node, initially it has four branches t1,t3,t4,t2But only task t4And t2The node is a node that is hopefully extended because t is completed at the user5After that, the task t has passed1And t3By a deadline, i.e. task t1And t3Has failed.
Step two: a scheduling algorithm based on a branch-and-bound idea is adopted, as shown in fig. 4, and the specific implementation steps are as follows:
(1) firstly, initializing, assigning a null value to a candidate task set of a root node, searching from the root node, wherein a task sequence is a null set, the candidate set is all tasks, and the score of the current optimal solution is 0.
(2) The search is then invoked recursively. Calculating a candidate task set and an upper bound of each node in the search tree; then sorting the task nodes in the candidate set in a descending order according to an upper bound; if the upper bound of the task in the candidate set is larger than the current optimal solution, continuing searching the node; otherwise, whether the value of the optimal solution is to be updated is checked.
(3) And finally, returning a task scheduling sequence corresponding to the optimal solution.
Step three: an approximate scheduling algorithm, namely the most promising branch heuristic algorithm, is adopted, and the specific implementation steps are as follows:
(1) the method comprises the steps of initializing, searching from a root node, setting a task sequence as a null set, setting all tasks in a candidate set, and setting the score of the current optimal solution to be 0.
(2) The search is then invoked recursively. Calculating a candidate task set and an upper bound of each node in the search tree; sorting the task nodes in the candidate set according to descending order of an upper bound, and selecting a branch node with the largest upper bound in each layer;
(3) and performing branch-bound scheduling algorithm search on the task node with the maximum upper bound in each layer, and returning a task scheduling sequence corresponding to the optimal solution.
Step four: an approximate scheduling algorithm, namely a heuristic algorithm based on width k search, is adopted, and the specific implementation steps are as follows:
(1) the method comprises the steps of initializing, searching from a root node, setting a task sequence as a null set, setting all tasks in a candidate set, and setting the score of the current optimal solution to be 0.
(2) The search is then invoked recursively. Calculating a candidate task set and an upper bound of each node in the search tree; sorting the task nodes in the candidate set according to descending order of an upper bound, and selecting k branch nodes with the largest upper bound in each layer;
(3) and performing branch-and-bound scheduling algorithm search on the k task nodes with the maximum upper bound in each layer, and returning a task scheduling sequence corresponding to the optimal solution.
For example, in the example of fig. 2, let k be 2. Firstly, starting from a root node, five task nodes t are respectively expanded in a first layer1,t2,t3,t4,t5And stores five task nodes, but only t1And t5With a larger upper limit, will t1And t5Storing; the second level nodes are then expanded to generate five task sequences, but only two of the best solutions (t)1,t4) And (t)5,t4) And storing; then continuing to expand the third layer and searching a leaf node t2By (t)5,t4,t2) Updating the current optimal solution to currMax 7; continue to search down to the fourth layer, retrieve the task sequence (t)1,t4,t5,t2) At this time, if currmax is 10, the task sequence is the optimal solution; by two task nodes t1And t5All the stored nodes are empty, and the search is stopped.
Claims (1)
1. A scheduling method of multi-type tasks in spatial crowdsourcing is characterized in that research objects are the multi-type tasks in a server task pushing mode, each task has characteristics, and a task execution path is planned for the tasks allocated to a user; the method comprises the following three aspects: a branch and bound scheduling method and two approximate scheduling methods;
1) calculating a candidate task set in multi-type task scheduling based on path planning;
2) the existing branch-and-bound thought algorithm is improved to adapt to the multi-type task scheduling environment, the characteristics of the multi-type tasks are fully combined, the task scheduling order is reasonably planned, and the algorithm can obtain an accurate execution path solution;
3) improving the operation efficiency on the basis of a branch-and-bound thought algorithm, and adopting an approximate scheduling algorithm, namely a most promising branch heuristic algorithm;
4) on the basis of the most promising branch heuristic algorithm, an approximate scheduling algorithm, namely a heuristic algorithm based on width k search, is adopted for improving the accuracy;
the step 1) further comprises the following steps:
1.1) calculating a candidate task set; firstly, acquiring a current task sequence, a candidate task set and a task to be expanded, and initializing the candidate set of the task to be expanded;
1.2) traversing all tasks in the current candidate set, checking whether the tasks in the candidate set can be completed within the deadline, and if the tasks can be completed within the deadline, adding the tasks into the next layer of candidate set; otherwise, the task can not be completed within the effective time, and the task is abandoned;
the step 2) further comprises the following steps:
2.1) adopting a scheduling algorithm based on a branch-and-bound thought; firstly, initializing, starting searching from a root node, wherein a task sequence is a null set, all tasks of a candidate set are collected, and the score of the current optimal solution is 0;
2.2) then recursively invoking the search; calculating a candidate task set and an upper bound of each node in the search tree; then sorting the task nodes in the candidate set in a descending order according to an upper bound; if the upper bound of the task in the candidate set is larger than the current optimal solution, continuing searching the node; otherwise, checking whether the value of the optimal solution needs to be updated;
2.3) finally returning a task scheduling sequence corresponding to the optimal solution;
the step 3) further comprises the following steps:
3.1) adopting a most promising branch heuristic algorithm; firstly, initializing, starting searching from a root node, wherein a task sequence is a null set, all tasks of a candidate set are collected, and the score of the current optimal solution is 0;
3.2) then recursively invoking the search; calculating a candidate task set and an upper bound of each node in the search tree; sorting the task nodes in the candidate set according to descending order of an upper bound, and selecting a branch node with the largest upper bound in each layer;
3.3) carrying out branch-and-bound scheduling algorithm search on the task node with the maximum upper bound in each layer, and returning a task scheduling sequence corresponding to the optimal solution;
the step 4) further comprises the following steps:
4.1) adopting a heuristic algorithm based on width k searching; firstly, initializing, starting searching from a root node, wherein a task sequence is a null set, all tasks of a candidate set are collected, and the score of the current optimal solution is 0;
4.2) then recursively invoking the search; calculating a candidate task set and an upper bound of each node in the search tree; sorting the task nodes in the candidate set according to descending order of an upper bound, and selecting k branch nodes with the largest upper bound in each layer;
4.3) carrying out branch-and-bound scheduling algorithm search on the k task nodes with the maximum upper bound in each layer, and returning a task scheduling sequence corresponding to the optimal solution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710595827.5A CN107317872B (en) | 2017-07-20 | 2017-07-20 | Scheduling method of multi-type tasks in space crowdsourcing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710595827.5A CN107317872B (en) | 2017-07-20 | 2017-07-20 | Scheduling method of multi-type tasks in space crowdsourcing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107317872A CN107317872A (en) | 2017-11-03 |
| CN107317872B true CN107317872B (en) | 2020-05-05 |
Family
ID=60178162
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710595827.5A Active CN107317872B (en) | 2017-07-20 | 2017-07-20 | Scheduling method of multi-type tasks in space crowdsourcing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107317872B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220780B (en) * | 2017-06-16 | 2020-07-28 | 佛山科学技术学院 | Heterogeneous task execution sequence optimization method in crowdsourcing system |
| CN108108390B (en) * | 2017-11-15 | 2019-02-19 | 北京达佳互联信息技术有限公司 | Data distributing method and device |
| CN109101329B (en) * | 2018-07-25 | 2021-02-12 | 陕西师范大学 | Fine-grained task allocation method and system for acquiring data through multiple mobile terminals |
| CN111461469B (en) * | 2019-01-18 | 2023-05-09 | 阿里巴巴集团控股有限公司 | Personnel scheduling method and computer equipment |
| CN112819285B (en) * | 2021-01-05 | 2023-09-26 | 北京百度网讯科技有限公司 | Flight matching processing and training method, device and equipment |
| CN113095618B (en) * | 2021-03-01 | 2024-04-02 | 华中科技大学 | IT operation and maintenance service personnel scheduling method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101604258A (en) * | 2009-07-10 | 2009-12-16 | 杭州电子科技大学 | A kind of method for scheduling task of embedded heterogeneous multiprocessor system |
| US7930198B2 (en) * | 2005-05-09 | 2011-04-19 | Siemens Corporation | Maintenance event planning and scheduling for gas turbines |
| CN102402722A (en) * | 2011-09-28 | 2012-04-04 | 上海交通大学 | Method for improving assembly line efficiency of project machine products |
| CN106055862A (en) * | 2016-05-05 | 2016-10-26 | 华东师范大学 | Novel efficient heuristic-type two-stage parallel branch-and-bound method |
| CN106779482A (en) * | 2017-02-17 | 2017-05-31 | 苏州大学 | The method for allocating tasks and system of dominance relation constraint between a kind of task based access control |
-
2017
- 2017-07-20 CN CN201710595827.5A patent/CN107317872B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7930198B2 (en) * | 2005-05-09 | 2011-04-19 | Siemens Corporation | Maintenance event planning and scheduling for gas turbines |
| CN101604258A (en) * | 2009-07-10 | 2009-12-16 | 杭州电子科技大学 | A kind of method for scheduling task of embedded heterogeneous multiprocessor system |
| CN102402722A (en) * | 2011-09-28 | 2012-04-04 | 上海交通大学 | Method for improving assembly line efficiency of project machine products |
| CN106055862A (en) * | 2016-05-05 | 2016-10-26 | 华东师范大学 | Novel efficient heuristic-type two-stage parallel branch-and-bound method |
| CN106779482A (en) * | 2017-02-17 | 2017-05-31 | 苏州大学 | The method for allocating tasks and system of dominance relation constraint between a kind of task based access control |
Non-Patent Citations (1)
| Title |
|---|
| 众包技术研究综述;冯剑红;《计算机学报》;20150930;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107317872A (en) | 2017-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107317872B (en) | Scheduling method of multi-type tasks in space crowdsourcing | |
| Qin et al. | Ride-hailing order dispatching at didi via reinforcement learning | |
| Li et al. | Truthful incentive mechanisms for geographical position conflicting mobile crowdsensing systems | |
| Xu et al. | A hybrid machine learning model for demand prediction of edge-computing-based bike-sharing system using Internet of Things | |
| CN105532030B (en) | For analyzing the devices, systems, and methods of the movement of target entity | |
| Samaan et al. | A mobility prediction architecture based on contextual knowledge and spatial conceptual maps | |
| Qu et al. | A cost-effective recommender system for taxi drivers | |
| CN109658033B (en) | Method, system, device and storage medium for calculating similarity of goods source route | |
| CN110462604A (en) | The data processing system and method for association internet device are used based on equipment | |
| CN111143680B (en) | Route recommendation method, system, electronic equipment and computer storage medium | |
| US20190012605A1 (en) | Artificial intelligence based solution generator | |
| CN110400128A (en) | A kind of space crowdsourcing method for allocating tasks based on the perception of worker's preference | |
| CN106651213A (en) | Processing method and device for service orders | |
| Xue et al. | Computational experiment-based evaluation on context-aware O2O service recommendation | |
| Liu et al. | Understanding taxi drivers’ routing choices from spatial and social traces | |
| CN107657357B (en) | Data processing method and device | |
| Chen et al. | Efficient dispatching for on-demand ride services: Systematic optimization via Monte-Carlo tree search | |
| WO2014124279A1 (en) | Customer experience management for an organization | |
| Widdows et al. | Grabshare: The construction of a realtime ridesharing service | |
| CN113780956B (en) | Logistics freight generation method, device, equipment and storage medium | |
| Tang et al. | The data and science behind grabshare carpooling | |
| Huang | A schedule-based pathfinding algorithm for transit networks using pattern first search | |
| WO2014107517A1 (en) | Priority-weighted quota cell selection to match a panelist to a market research project | |
| Zheng et al. | Multi-campaign oriented spatial crowdsourcing | |
| Vateekul et al. | Recurrent Neural-Based Vehicle Demand Forecasting and Relocation Optimization for Car-Sharing System: A Real Use Case in Thailand |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |