| Constrained layout problem studying has a wide application background,for example: layout design of satellite cabin,layout of factory and equipment,large-scale equipment manufacturing and plate cutting of steel enterprises,etc.In this paper,the loading and packing design of multi-module satellite is studied,and the loading is simplified to cylinder and cuboid.This is belonged to NP-hard problem,the solving has time complexity with exponential order,the common methods are heuristics and evolutionary algorithms at present.However,these methods didn't take the multimodule configuration optimization problem of loading into account before packing.Otherwise,the interference calculation of evolutionary solution based on stochastic initial scheme is time-consuming.The objective function of optimization model of multi-module integrated loading scheme is high dimension and multi-objective conflicting,which is hard to obtain the optimum solution.Therefore,the best solution method at present is to allow inter-cabin migration in multi-module loading optimization by decomposing the high-dimensional solution space and optimizing the multi-objective grouping mechanism.These mechanisms can only reduce the complexity of problem solving to a certain extent.Inspired by the no-free lunch theorem and previous work,this paper studies the heuristic ant colony optimization method for multi-module satellite payload allocation and constrained single-cabin loading,including knowledge-based load allocation and regional heuristic ant colony optimization and multi-cabin loading particle swarm collaborative optimization.The main innovation includes two aspects:1.A knowledge-based method for optimal load allocation and verification is proposed.Firstly,the knowledge of payload multi-module allocation is acquired based on equilibrium mechanics and inertia definition.Then,based on the objective function,integrate the Knowledge into heuristic ant colony and iterate,and search the optimum scheme for multi-cabin allocation of load.Otherwise,based on the acquired knowledge,load checks are carried out at each iteration of each cabin loading,and unreasonable load allocation are transferred among the cabins.The experimental data shows that the proposed load optimal allocation and verification method can effectively reduce the unreasonable load multi-module allocation.2.A heuristic swarm Intelligence optimization method of knowledge-based regional location and ant colony optimization ordering is proposed.By dynamically generating the rectangular area filled with loads,and based on the greedy strategy,searching the optimal location and direction of loads in the generated region set,this method combines the knowledge and ant colony iteration and search the loading sequence of each cabin.Because of the compactness and feasibility of the loading scheme which this mechanism constructs,and there is no large time-consuming interference calculation,the proposed performance is obviously improved.The experimental data show that the loading method proposed in this paper reduces more moment of inertia,envelope radius and calculation time of the whole scheme and has better stability than the best reported method.This paper takes the multi-module loading design of commercial satellite as an example,studied a heuristic ant colony optimization method of multi-module load allocation and constrained loading of satellite.By theoretical analysis and experimental observation,the knowledge of satellite loading design is acquired,it is combined with swarm intelligence organically,and intelligent load allocation and filling optimization are realized.However,the dynamic generation of the region and the load location and orientation strategy ensure that the loading scheme is compact and feasible,and huge interference calculation is avoid,thus the performance of the algorithm is obviously improved.It is hoped that this method can provide references for solving other complex layout design problems. |
PDF Full Text Request