Role-oriented Task Allocation in Human-Machine Collaboration System

In the past few years, collaborative AI-infused machines have been introduced as a new generation of industrial “workers”, working with humans to share the workload. These “workers” have the potential to realize Human-Machine Collaboration (HMC),which enables flexible automation. However, combining intelligent machines with humans to obtain more efficient and accuracy human-in-the-loop solutions is a nontrivial task. Therefore, how to allocate tasks between humans and machines has become an important issue in system design. Inspiring by the graph path searching, in this paper, we adopt an acyclic direction graph to construct the role-oriented task allocation problem, and develop an Ant colony optimization based Human-Machine Task Allocation (A-HMTA) approach to find an optimized allocation solution in the search space. Experimental results show that our approach is superior to traditional approaches in terms of cost and time consumption.