Research Of Optimization Of Slab Loading Problem

Slab manufacture of a steel company has witnessed a rapid growth in recent years, while the importance of the warehouse optimization is booming up as well. Usually, a warehouse contains hundreds of stacks as well as thousands of slabs. Moreover, each stack and each slab has a large quantity of information and data. If the slabs are stacked by the worker on site, it will not only lead to a terribly hard work, but also poor utilization of the space.Our research can be divided into three parts: modeling, algorithm and simulation.1. From the point of view of modeling, a mathematical model is established after studying on a lot of papers as well as communication with the works on site.2. From the point of view of algorithm, the main work of our research is stated as the following:First of all, we start from the existing Weighted Best Fit (WBF) algorithm. Based on this algorithm, genetic algorithm is adopted to optimize the slab sequence. Then, we use variation function and non-uniform mutation to get rid of premature convergence, thereby forming a kind of algorithm mixed by WBF and genetic algorithm. Moreover, we raise a multi-objective optimization problem (MOP), which is much closer to the reality. In this part, according to the original PESAII algorithm, we adopt several strategies to overcome the defect of the original PESAII algorithm, whose solution set is short of diversity. The strategies are randomized weight, weighted hypebox, external population variation function and random generation of internal population.Last but not least, we also consider about an extremely bad condition that the warehouse is about to saturate. Some of the slabs can't be put into the stack. Our solution is to utilize the fuzzy inference and judgment to make decision that which constraints should be broken to load the slabs, resulting in the least cost and risk.3. From the point of view of simulation, this dissertation develops a slab loading optimization application, which is useful for both research based on simulation data and practical use.