Hierarchical Pricing Mechanism With Financial Stability for Decentralized Crowdsourcing: A Smart Contract Approach

Software crowdsourcing is an emerging approach to software engineering with great potential for the subdivision and assignment of large-scale tasks. However, because of the centralization of the traditional crowdsourcing platform, information disclosure and nontransparent accounting may be difficult to avoid. To address this issue, we first introduce a novel blockchain-enabled crowdsourcing platform that integrates the functions of task assignment and resource lending via two dedicated smart contracts. Second, to ensure financial stability in the blockchain-enabled market and to match the difficulty of the received tasks with the ability of the workers, we design a dynamic, hierarchical pricing mechanism based on economic modeling methods and heterogeneous agent theory. With this mechanism, the market is divided dynamically into multiple levels according to the remuneration of the customers' offer and the market value of the workers' resources. Additional constraints are proposed to avoid possible malicious trading behavior from workers in the resource lending process. We prove theoretically the rationality of our model and demonstrate the dynamics of the model. We show that the market price and demand can be convergent and test the cost of executing the two smart contracts. Finally, extensive experimental results demonstrate the correctness and feasibility of the platform and confirm that the hierarchical pricing mechanism can maintain the stability of the market.

I. Introduction

As an emerging software engineering approach, crowdsourcing is making a significant impact [1], [2]. As shown in Fig. 1, the two main roles in most traditional crowdsourcing platforms are those of customers or workers. With a crowdsourcing platform, the customers publish tasks they wish to be undertaken and the workers can pick tasks that match their abilities. All task assignments are made through the platform. If existing resources (such as a training data set or computing resources) are inadequate, workers may purchase resources from third-party resource providers. This type of crowdsourcing platform has several advantages. For the customers, the crowdsourcing platform can help them save on development budgets and time [3]. They may also discover more innovative solutions [4]. For the workers, the crowdsourcing platform not only provides them with an income channel but can also help improve their skills through practice. Fig. 1.

Centralized model of the traditional crowdsourcing platform.