Solving buffer allocation problem in production lines using tabu search based approaches

The buffer allocation problem, which involves the distribution of buffer space among the intermediate buffers of a production line, arises in a wide range of manufacturing systems, and it is one of the most important optimization problems faced by manufacturing systems designers. The primary aim of this Ph.D. study is to introduce novel tabu search based solution approaches for solving buffer allocation problem in production lines. In this thesis, the buffer allocation problem is solved in three stages. In the first stage, a novel TS algorithm including new move definitions is proposed to solve the buffer allocation problem under the objective of throughput maximization for homogeneous production lines involving unreliable machines with deterministic processing times. Following a pilot experiment to identify the best TS parameters, the new move definitions for buffer allocation problem are introduced. In the second stage, the problem is extended to non-homogeneous production lines, and an adaptive TS algorithm is proposed to solve the revised problem under the objective of throughput maximization. Besides proposing a new strategy to tune the parameters of TS adaptively during the search, an experimental study is carried out to select an intelligent initial solution scheme among three alternatives so as to decrease the search effort to obtain the best solutions. Finally, in the last stage, three approaches are proposed to solve the buffer allocation problem for non-homogeneous production lines involving unreliable machines with deterministic processing times. These three approaches which integrate binary search, tabu search, and simulated annealing with an adaptive tabu search mechanism aim at minimizing the total buffer size to achieve a desired throughput level. To improve the searching efficiency of TS and SA algorithms alternative neighborhood generation mechanisms are suggested and their performance are tested.