A truthful double auction for two-sided heterogeneous mobile crowdsensing markets
Introduction
Mobile Crowdsensing (MCS) [1] has become a promising paradigm to solve large-scale and complex data collection problems by leveraging pervasive sensor-equipped mobile devices of the crowd (e.g., smartphones, Tablet PCs, wearable devices). In a MCS platform, a big data collection problem is usually divided into small tasks, thus mobile device users can contribute their relatively small sensing capacities to accomplish the problem in a crowdsourcing way. So far, a number of crowd sensing applications have been developed for a variety of purposes, such as traffic information collecting [2], environment monitoring [3], crowdsourced commercial activities [4], and so on.
One critical issue in practical MCS applications is how to stimulate users to participate in the sensing campaigns, as executing sensing tasks will undoubtedly consume physical resources of devices, time, and even human intelligence. Similar to the employment relationship in a labor market, appropriate incentives are needed to compensate users in order to recruit enough sensing workforce.
However, current incentive mechanisms are mainly based on a monopoly campaign scenario, which means they simply assume all sensing tasks are belong to a globally unique campaign, and all participating users are working for the same campaign. This may be reasonable for some specific-purpose applications, but things become different in an integrated MCS platform. In such a platform, many kinds of MCS applications are available, and different requesters can publish multiple sensing campaigns in a unified platform concurrently. Some platforms such as Medusa [5] and gPS [6] have been designed in this way. Integrated MCS brings significant benefits, including total overhead reduction and more efficient management. What is more, it also relieves the users from switching between various MCS platforms and provides them more sensing choices.
In this paper, we consider the case where multiple campaign requesters demand for sensing workforce and multiple users supply their sensing capacities. Besides, both requesters and users are self-interested and want to maximize their own benefits strategically. This actually builds up a further developed two-sided market, which is fundamentally different from the traditional monopoly campaign scenario. As multiple requesters with different interests may compete for potential sensing capacities, it is unsuitable to regard all campaigns as a whole and apply a single incentive mechanism in a monopolistic way. Letting each campaign apply an independent incentive mechanism is also inconvenient, since users have to contemplate different strategies among various campaigns to maximize their utilities. Indeed, none of existing incentive mechanisms [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19] can work well in the multiple-campaign competing situations.
In order to fairly and effectively allocate resources and facilitate trades between requesters and users, we propose a Truthful Double auction mechanism for the two-sided Mobile Crowdsensing market (TDMC). However, the design is challenging as the market is intrinsically heterogeneous. Further to say, requesters demand for tasks with diverse requirements in aspects of locations, time and sensing methods, while users have different availabilities and preferences for different tasks based on their spatio-temporal and device states. This makes it extremely difficult to allocate tasks and achieve efficiency, i.e., the total valuation of all allocated resources of all requesters and users, or called the social welfare, is maximum. Besides, truthfulness is critical for an auction, which means bidders should truthfully reveal their private information for the items they bid. However, to fully stimulate the sensing capacities, users should be allowed to freely bid for multiple heterogeneous tasks as long as they can undertake. Their private information is actually multi-dimensional, which makes it harder to ensure truthfulness. What is more, it is also important for a double auction to guarantee: (i) individual rationality: bidders have non-negative utilities by reporting truthfully, (ii) budget balance: auctioneer would not suffer a deficit, and (iii) computational tractability: auction runs in polynomial time.
To overcome above challenges, we first categorize all tasks from different campaigns into orthogonal sensing patterns to characterize the heterogeneity of the market. Then we allow users to bid personalized maximum available workload and associated (unit) costs for different patterns based on their sensing capacities. Finally, we adopt a two-stage allocation approach to determine the trading results and optimize the social welfare. In stage one, by introducing a carefully designed virtual padding requester, we intentionally intensify the competition among requesters and screen out a set of more competitive requesters to be the winners. In stage two, we match these selected requesters with users who offer the cheapest workload and get the final allocation. Novel pricing schemes are also designed for both requesters and users, which ensures the truthfulness as well as budget balance. Furthermore, we also show TDMC can asymptotically approach the efficiency as the workload supply compared with demand becomes more and more sufficient. The contributions of this paper are as follows.
- •
To the best of our knowledge, TDMC is the first truthful auction mechanism for a two-sided heterogeneous MCS market, where multiple requesters and users have diverse demands and supplies. A joint consideration of competition among requesters and heterogeneity in the market significantly complicates the design.
- •
Utilizing a padding idea, TDMC develops a two-stage allocation approach and corresponding pricing schemes for requesters and users to achieve approximate efficiency while preserving truthfulness and budget balance.
- •
We theoretically prove that TDMC possesses the attractive properties of truthfulness, individual rationality, budget balance, computational tractability, and asymptotic efficiency as the workload supply compared with demand becomes more and more sufficient.
- •
We further consider several practical issues to make TDMC more adaptable, including more flexible bid profiles for both requesters and users, and two adjustment methods of price intervention and workload quota to control the sensing quality.
The rest of the paper is organized as follows. In Section 2, we review the related work. In Section 3, we introduce the model and formulate the problem. In Section 4, we elaborate the details of TDMC and provide theoretical proofs of desired properties. Then in Section 5 we consider several practical issues of TDMC. In Section 6 we show some simulation results. In Section 7, we conclude the paper.
Section snippets
Incentive mechanisms for mobile crowdsensing
A number of incentive mechanisms [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19] have been developed for MCS. In [8], [9], the platform (also the campaign organizer) applies a stackelberg game to maximize its utility by deciding an optimal total reward. Koutsopoulos [7] proposes a reverse auction in which the platform determines users’ participation levels based on their reported unit costs to minimize the total payments under certain service quality limits. An all-pay
System model and problem formulation
In this section, we first give some preliminaries and describe the system model for a two-sided heterogeneous MCS market. Then we formulate the problem and state our design targets.
Mechanism details and analysis
In this section, we first elaborate the details of TDMC. Then, we prove TDMC is truthful, individual rational, budget balanced, computational tractable, and asymptotic efficient as the workload supply compared with demand becomes more and more sufficient.
Practical issues
In this section, we further consider several practical issues to make TDMC adaptable in more general cases, including two more flexible bid profiles for both requesters and users, and two adjustment methods to control the sensing quality. Notably, all properties proved in Section 4.2 are still preserved. Besides, we also discuss the privacy and collusion issues briefly.
Simulation results
In this section, we conduct extensive simulations to evaluate the performance of TDMC. All simulations are run on MATLAB. We first implement TDMC and its extended form named TDMC-D which considers the discounted bid profiles for requests. For comparison, we introduce two mechanisms denoted by Uniq-one and Uniq-all, which means they only allow each user to bid a unique cost for all tasks. In Uniq-one, users only bid for maximum available workload of a single pattern by reporting corresponding
Conclusion
In this paper, we propose a truthful double auction TDMC for two-sided heterogeneous mobile crowdsensing markets. TDMC develops a two-stage allocation approach and corresponding pricing schemes for both requesters and users. We prove that TDMC is truthful, individual rational, budget balanced, computational tractable, and asymptotic efficient as the workload supply compared with demand becomes more and more sufficient. Several practical issues are also considered to make TDMC more adaptable.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (Nos. 61132001, 61120106008, 61472402, 61472404, 61272474, 61202410, 61502457 and 61572476).
References (37)
- et al.
Heterogeneous-belief based incentive schemes for crowd sensing in mobile social networks
J. Netw. Comput. Appl.
(2014) A dominant strategy double auction
J. Econ. Theory
(1992)- et al.
Preserving privacy in participatory sensing systems
Comput. Commun.
(2010) - et al.
Efficient and privacy-preserving data aggregation in mobile sensing
Proceedings of the 20th IEEE International Conference on Network Protocols (ICNP)
(2012) - et al.
A survey on privacy in mobile participatory sensing applications
J. Syst. Softw.
(2011) - et al.
Mobile crowdsensing: current state and future challenges
IEEE Commun. Mag.
(2011) - et al.
Vtrack: accurate, energy-aware road traffic delay estimation using mobile phones
Proceedings of the 7th ACM Conference on Embedded Networked Sensor Systems (SenSys)
(2009) - et al.
Common sense: participatory urban sensing using a network of handheld air quality monitors
Proceedings of the 7th ACM Conference on Embedded Networked Sensor Systems (SenSys)
(2009) - et al.
Automatic collection of fuel prices from a network of mobile cameras
Proceedings of the 4th IEEE International Conference on Distributed Computing in Sensor Systems (DCOSS)
(2008) - et al.
Medusa: a programming framework for crowd-sensing applications
Proceedings of the 10th ACM International Conference on Mobile Systems, Applications, and Services (MobiSys)
(2012)
A generic participatory sensing framework for multi-modal datasets
Proceedings of the 9th IEEE International Conference on Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP)
Optimal incentive-driven design of participatory sensing systems
Proceedings of the 32th IEEE International Conference on Computer Communications (INFOCOM)
Crowdsourcing to smartphones: incentive mechanism design for mobile phone sensing
Proceedings of the 18th ACM Annual International Conference on Mobile Computing and Networking (MobiCom)
Incentive mechanisms for smartphone collaboration in data acquisition and distributed computing
Proceedings of the 31th IEEE International Conference on Computer Communications (INFOCOM)
Profit-maximizing incentive for participatory sensing
Proceedings of the 33th IEEE International Conference on Computer Communications (INFOCOM)
Crowdsourcing with tullock contests: a new perspective
Proceedings of the 34th IEEE International Conference on Computer Communications (INFOCOM)
How to crowdsource tasks truthfully without sacrificing utility: online incentive mechanisms with budget constraint
Proceedings of the 33th IEEE International Conference on Computer Communications (INFOCOM)
Trac: truthful auction for location-aware collaborative sensing in mobile crowdsourcing
Proceedings of the 33th IEEE International Conference on Computer Communications (INFOCOM)
Cited by (15)
Truthful online double auctions for on-demand integrated ride-sourcing platforms
2023, European Journal of Operational ResearchA collaboration-driven mechanism for AI diagnose with multiple requesters under incomplete information
2023, Computer NetworksCitation Excerpt :Therefore, some researchers take multiple requesters into consideration [9,10]. Chen and Liu use truthful double auction to consider mutual competition [11] where different requesters always compete for potential MCS task performers. Competition may lead to several problems, and the main problem is, in competition, requesters with the highest bidder get the most accurate data, while other requesters just have to settle for inaccurate data (inaccurate data is a combination of spurious data and incomplete data) or no data.
User satisfaction aware maximum utility task assignment in mobile crowdsensing
2020, Computer NetworksCitation Excerpt :In such systems, the workers should be compensated financially by the crowdsourcer for their efforts and disbursements. In order to achieve this, various incentive mechanisms (e.g., auctions, lottery contests, reputation based models) have been proposed [25], among which auction models are the most popular and commonly used [26–28]. In these models, generally, workers estimate and announce the level of effort they will have to put in to complete each task, and the platform strategically selects the workers and determines how much they will be rewarded while ensuring the quality of the sensed data.
Incentives-based preferences and mobility-aware task assignment in participatory sensing systems
2018, Computer CommunicationsCitation Excerpt :This induces a loss of tasks and a low data-quality especially in terms of temporal accuracy. Furthermore, current incentives mechanisms [8,9,13–20] are usually determined “a priori” and not adapted to the progress of the assignment process in terms of users’ resources or data samples’ quality which leads to unwilling users to participate to crowdsensing. To address these issues, we consider, in this paper, an additional distributed crowdsensing assignment phase as introduced in our previous work [21].
Double Auction based Incentive Mechanism for Participatory Sensing Systems
2023, Research Square