Evolutionary Fuzzy‐based gravitational search algorithm for query optimization in crowdsourcing system to minimize cost and latency. (23rd October 2020)

• Record Type: Journal Article
• Authors: Bhaskar, N.; Kumar, P. Mohan; Renjit, J. Arokia
• Is Part Of: Computational intelligence. Volume 37:Number 1(2021)
• Page Start: 2
• Page End: 20
• Publication Date: 2020-10-23
• Subjects: cost -- crowdsourcing -- evolutionary fuzzy‐based gravitational search algorithm -- latency -- query processing -- top k‐query; Artificial intelligence -- Periodicals; Computational linguistics -- Periodicals; 006.3
• DOI: 10.1111/coin.12382
• Languages: English
• ISSNs: 0824-7935

Abstract

Abstract: Crowdsourcing is an environment where a group of users collaborates together to exchange information and to find answers for complex problems (queries). Query optimization is the task of selecting the best query strategy with less cost associated with it. The crowdsourcing cost can be determined by selecting the best plan from the set of options available and the best plan considerably reduce the cost for the inquiry configuration. As one of the center tasks in information recovery, the investigation of top‐k queries with crowdsourcing, to be specific group empowered top k inquiries is depicted. This issue is defined with three key variables, latency, money related expense, and nature of answers. The fundamental point is to plan a novel system that limits financial cost when the latency is compelled. In this article, we used a heuristic search algorithm named as Evolutionary Fuzzy‐based Gravitational Search algorithm (EFGSA) that produces an optimal query feature selection results with minimizing cost and latency. EFGSA‐based crowdsourcing framework gives a better balance between latency and cost while generating query plans. The performance analysis of proposed EFSGA for optimal query plan is evaluated in terms of running time, accuracy, monetary cost, and so on. From the experimental results, the proposed method achieved better results than other methods in our cost and latency model.