Research On Check-in Baggage Loading Algorithm With Complicated Constraints

As one of the key tasks for airport,achieving full-automatic baggage handling is of great significance to the construction of "smart airport".In this paper,algorithms are designed to solve the problems of labor-intensive and inefficient in the process of baggage trailer loading and aircraft compartment loading.Through the summary of its academic connotation "three-dimensional bin-packing problem with strong heterogeneous",it is found that the constraints,optimization objectives and actual requirements of the existing algorithms are not matched well.For this reason,the clustering algorithm in machine learning and special space constraints in aircraft compartment are introduced.Two kinds of suitable bin-packing problems are determined,and solutions are put forward.Works are depicted in detail as follow:Mathematical models of two kinds of bin-packing problems are constructed.Based on the analysis of baggage trailer loading and aircraft compartment loading,a multi-box model with trailer constraints and a single-box model with compartment constraints are constructed.The model is expressed in mathematical language to support the algorithm design.Aiming at the problem of baggage trailer loading,an efficient algorithm with clustering pretreatment process is designed.It combines K-means clustering with wall-building heuristic algorithm,divides a strong heterogeneous problem into several weak heterogeneous problems,and solves them one by one.Experiments with passenger baggage examples,the rationality of the algorithm is verified.Testing with international examples(BR1~BR7),the outstanding advantage of the solution time is proved 62.3% lower than that of similar algorithms.Moreover,the introduction of K-means clustering makes the algorithm have an excellent characteristic that the that the computing time is not easily affected by heterogeneity.Aiming at the problem of aircraft compartment loading,an algorithm with the special space constraints of compartment is designed.Based on the idea of "key points",it determines the location of baggage one by one through dimension reduction and dimension elevation.In order to verify the effectiveness of algorithm,three groups of passenger baggage,which are larger,smaller and hybrid,are used to experiment.It is proved that the algorithm has the ability to complete the task of loading baggage quickly under various operating conditions.In this paper,machine learning algorithm and irregular space constraints are introduced into traditional bin-packing problem,which not only makes up for the shortcomings of existing algorithms applied to check-in baggage loading,but also opens up a new way to solve the bin-packing problem under the conditions of multiple space constraints,strong heterogeneity of goods and high sensitivity of computing time.