Orantılı esnek açık atölye tipi çizelgeleme için karınca kolonisi optimizasyonu yaklaşımı

Shop scheduling problems have exceptionally wide application fields both in manufacturing and service sectors. Multiprocessor open shop is among common shop environments and it consists of at least two machine centers with one or more center having parallel machines for the same task. There are n jobs to visit the centers without a predefined route. The shop widely exists particularly in diagnostic medical testing, repair and maintenance services, inspection and quality control operations and electronics manufacturing processes. However, the problem gained little attention in the literature. There has been an increase in the number of researches in the field in recent years but still there is considerable room for improvement. In this thesis study, the proportionate multiprocessor open shop problem was considered where proportionate feature refers to processing times of machine centers being fixed and independent of the job. An Ant Colony Optimization algorithm was proposed for the problem. The algorithm is based on a very efficient novel solution representation of the problem. The proposed algorithm further employs random exploration and local exploration (analogous to local search) routines. Exploitation of search knowledge and problem-specific knowledge was incorporated with tailored uses of pheromone information and heuristic information, respectively. The algorithm was tested on 100 benchmark instances from the literature. Comparisons showed that it outperformed the current state-of-the-art scatter search with path relinking algorithm both in solution quality and computational time. Of particular importance is its performance in large-scale instances and the relatively short time it required to reach the high-quality results.