Controlled parameters in suboptimal solution

In this paper, two results are presented. Firstly, we discuss the role of a controlled parameter in the solution process of using heuristic methods to obtain a suboptimal solution for an optimization problem. We find that using qualitative analysis associated with simulations can estimate the increase or decrease of the objective value with respect to the variation of the parameter and provide an information whether to vary the parameter to achieve a better suboptimal solution. Secondly, to overcome computational complexity, we derive a sectional processing algorithm for long-horizon job-shop-flow scheduling problem. We found that the objective value is improved after we decomposed the long-horizon into consecutive short horizons. This fact reflects the first result. We have simulated the proposed sectional processing algorithm on quite a few examplses and obtain very good results in the aspects of CPU times and objective values.