Rescheduling under disruptions in manufacturing systems. Models and algorithms (Q2175438)

The book is a volume out of a series of books handling ``Uncertainty and Operations Research''. The authors are from China and the UK and all their work is not only focused on solving problems, but also on adding computational experience for all the solution methods. The small book (about 140 pages) is organized in 5 problem-solving parts and an introduction part introducing the scheduling notations and the complexity status. The first model is formulated for \(m\) parallel machines and to minimize the mean completion times. This problem is easy to solve without any other restrictions. Here an unexpected machine breakdown is considered, which changes everything: The problem becomes NP-hard (a reduction to 3-partition is shown) and the previously optimal solution is no longer optimal, sometimes even no longer feasible. For this extension, they provide a pseudo-polynomial approach and give numerical experience for test-data. The next model treats the same base-problem with the extension that some jobs are not available at time. For this also NP-hard problem they present a pseudo-polynomial algorithm called ``Branch-and-price algorithm''. As before, an extensive practical part with numerical test-data is added. The third model includes newly arrived jobs during processing and the fourth-model also includes maintenance phases during processing. As before, the authors provide solution algorithms and test them on practical data. The last treated model reconsiders machine breakdowns and deterioration effects. Here they provide a ``Knowledge-based evolutionary'' approach for planning in advance. As always in the book, they add numerical practical experience at the end of the section. Conclusion: The book presents an interesting inside to problems when scheduling is disturbed by breakdowns, unavailable jobs or arriving jobs during processing possibly including maintenance phases. All problems are NP-hard and are solved individually with an assortment of different ideas and all models are tested numerically in practice.