TCBB Scheme: Applications to Single Machine Job Sequencing Problems

Transpose-and-Cache Branch-and-Bound (TCBB) has shown promise in solving large single machine quadratic penalty problems. There exist other classes of single machine job sequencing problems which are of more practical importance and which are also of considerable interest in the area of AI search. In the weighted earliness tardiness problem (WET), the best known heuristic estimate is not consistent; this is contrary to the general belief about relaxation-based heuristic. In the quadratic penalty problem involving setup times (SQP) of jobs, the evaluation function is non-order-preserving In this paper, we present the TCBB scheme to solve these problems as well. Experiments indicate that (i) for the WET problem, the TCBB scheme is highly effective in solving large problem instances and (ii) for the SQP problem, it can solve larger instances than algorithm GREC in a given available memory.

• Type: Conference
• Year: 2000
• Where: AAAI
• Authors: Sakib A. Mondal, Anup K. Sen
• Keywords: AAAI 2000, Intelligent Agents, Machine Quadratic Penalty, Quadratic Penalty Problem, Single Machine