Application of Improved Q Learning Algorithm to Job Shop Problem

王超,郭静,包振强
DOI: https://doi.org/10.3724/sp.j.1087.2008.03268
2008-01-01
Abstract:The Job Shop Problem (JSP) is a key technology in manufacturing system, and the Q learning was used to realize it. The improved Q learning algorithm was suggested because of the traditional algorithm has limitations of slow and partial constringency. In this algorithm, a complex action group was suggested. And uniformity degrees of action choices were weighed by cohesion, and then the two limitations could be overcome effectively. It could be adapted to the complicated manufacturing environment. Experimental results show its good effect in the JSP.
What problem does this paper attempt to address?