Many Is Better Than All: Efficient Selfish Load Balancing in Mobile Crowdsourcing Systems

In this work, we investigate the selfish load balancing problem in mobile distributed crowdsourcing networks. Conventional methods heavily relied on cooperations among users to achieve balanced resource utilization in platform-centric view. In achieving fairly low communication and computational overhead, and maintaining good load balancing property among selfish users, we resort to the ´d-choice´ method based on ´Ball and Bin´ theory [1] for balancing with limited information, and proliferate the ´Proportional Allocation´ [2] scheme for selfish load balancing. We combine the good properties in aforementioned schemes and propose ´Chance-Choice´, a lightweight distributed load balancing scheme for selfish users with fast convergence property. We find that, even with limited information, the balancing performance could be improved significantly, under the rule of opportunistic offloading and selfish behavior. Extensive evaluations have been made to show that, ´Chance-Choice´ outperforms several existing algorithms. Typically, comparing with ´Proportional Allocation´ scheme [2], ours could decrease the load gap by 50% to 80%, and reduce the overhead complexity from O(n) to O(1) comparing with the ´Max-weight Best Response´ algorithm [3], where n denotes the number of mobile users in crowdsourcing system.