改进的 memetic 算法在产品装配序列规划中的应用

12 浏览量更新于2024-08-28 收藏 2.35MB PDF 举报

本文主要探讨了在产品设计和制造过程中应用遗传算法（Memetic Algorithm, MA）来解决组装序列规划（Assembly Sequence Planning, ASP）问题。ASP是一项关键任务，其目标是确定组件的装配顺序，以优化生产成本和时间效率。然而，随着产品组件数量的增加，这个问题变得越来越复杂，传统的基于图形的方法往往难以提供有效的解决方案。遗传算法是一种混合智能搜索方法，它结合了遗传学中的自然选择、交叉和变异等概念，以及局部搜索策略，以求得全局最优解或近似最优解。在ASP中，MA的应用可以有效地处理大规模组合优化问题，通过迭代过程不断改进解的质量，避免陷入局部最优，从而找到更高效的组装路径。文章首先概述了ASP的重要性以及传统方法的局限性，然后介绍了MA的基本原理和优势，特别是在解决复杂问题时展现出的适应性和进化能力。MA通常包括以下几个关键步骤： 1. 初始化种群：创建一组随机的初始组装序列作为个体，每个序列代表一个可能的解决方案。 2. 适应度评估：根据某种评价函数，如总组装时间、所需的工位数或成本，计算每个序列的适应度值。 3. 选择操作：根据适应度值选择部分个体作为父代，进入下一轮迭代。 4. 变异操作：对父代进行变异，产生新的个体，这有助于跳出局部最优。 5. 交叉操作：将两个或多个父代个体的部分基因进行交换，生成新的可能的组装序列。 6. 局部搜索：在每次迭代后，使用局部搜索策略进一步优化找到的解决方案。 7. 重复迭代：上述步骤不断循环，直到达到预设的停止条件，如达到最大迭代次数或适应度值不再显著提高。通过这种方式，MA能够处理ASP问题中复杂的依赖关系和约束条件，提高了优化效率。研究者们在2009年7月提交并接受了这篇文章，表明了他们对于在ASP领域中引入MA作为新型优化工具的兴趣和信心。总结来说，这篇论文是关于如何利用遗传算法这一强大的优化工具来提升组装序列规划的性能，特别是在面临大量组件和复杂约束的现代工业环境中。MA的应用不仅有助于降低成本和时间，还能推动制造业向更高效率和智能化方向发展。然而，需要注意的是，该研究的结果必须在遵守版权法和非商业使用条款的前提下进行传播和讨论。

Homem et al. [7] applied the cut-set algorithm to

generate the assembly sequence. It can be assumed as a

traditional geometric feasibility method. The method was

based on the principle that disassembly sequence was the

inverse of the assembly sequence. There were two main

steps in this method: firstly, a liaison graph was built

according to the assembly model of products, and then a

cut-set algorithm was used to conduct an extensive search

to generate disassembly and/or graph that contained all

feasible disassembly sequences. The assembly sequence

can be obtai ned by inversing the disassembly sequence.

Based on traditional geometric feasibility reasoning ap-

proach, Su [8] proposed a systematical approach called

geometric constraint analysis, which uses the constraint

direction and the minimal constraint assembly state to

reason the assembly precedence and get the optimal

assembly sequence. Dong et al. [ 9 ] proposed a method

based on the geometry reasoning with the knowledge of the

directed-connectors after analyzing the characteristics of

directed-connectors, and presented a uniform representation

of parts. According to their research, the approac h could

reduce the computation complexity drastically and the

obtained assembly sequences were more feasible and

practical.

The knowledge-based approach applied the knowledge

of the existing assembly sequence and the past experi-

ence. The knowledge w as usually the assembly sequence

of the typical products in practice. Based on the concept

of the connector proposed by Tseng et al. [10]in1999,

Yin et al. [11]proposedanapproachtofindfeasibleand

practical plans for mechanical assemblies based on a

connector-based structure (CBS) hierarchy, which is a

knowledge-based approach. They proposed an efficient

representation of assembly inform ation to support the

identification of CBSs a nd presented algorithms to bui ld

a CBS hierarchy for an assembly from its connector-

based relat ional mod el auto matically. It is based on the

idea of reusing the existing plan and modifying the

existing plan and geometric reasoning. Dong et al. [12 ]

proposed a collaborative approach to ASP, which is also a

knowledge-based approach. They applied a role-based

model to compress or simplify the product and assigned

the different planning tasks to different designers who

carried out the collaborative planning respectively. Dong

et al. [13] proposed the conn ection-semantics-based

assembly tree (CSBAT) that provi des an appropria te

waytoconsiderbothgeometric information and non-

geometric knowledge. The ways to construct plans for

CSBAT were presented by retrieving the typical base, by

retrieving the standard base or by geometric reasoning.

Zha et al. [14] presented a n ew methodology to generate

all f easible as sembly seque nces of the pro duct by

reasoning and decomposing the feasible subassemblies,

and representing them by using the assembly Petri net

modeling.

In recent years, a lot of researchers began to use heuristic

algorithm to solve the ASP. Zhou et al. [ 15] used a genetic

algorithm to solve the assembly sequence planning prob-

lem. When initializing the population, the expert knowl-

edge was used to improve the composition of population by

inputting the feasible assemble sequence. Guan et al. [16]

used a gene-group-based evolution approach to optimize

assembly processes rather than merely assembly sequences.

A compound c hromosome encode was constructed to

represent th e ab undant assembly process info rmation.

Several specific genetic operators were used to generate

offspring individuals, and an integrated interference matrix

was built to determine whether an assembly process was

feasible geometrically. Li et al. [17] combined the strength

of the GAs and tabu search and proposed a tabu-enhanced

genetic algorithm to optimize the assembly sequence, while

the tabu search is used as local search. Chen et al. [18]

developed a three-stage integrated approach for assembly

sequence planning by using neural networks. In the first

stage, Above Graph and transforming rule were used to

create a correct explosion graph of the assembly models. In

the second stage, a three-level relational model with

geometric constraints and assembly precedence diagrams

were generated to create a compl ete relational model graph

and an incidence matrix. In the third stage, the back-

propagation neural network was employed to optimize the

available assembly sequence. Zhou et al. [19] applied a

modified genetic simulated annealing algorithm to optimize

the assembly sequence. Wang et al. [20] applied a novel ant

colony algorithm to solve the assembly sequence problem.

The local pheromone updating rule and the global phero-

mone updating rule were introduc ed. T he f irst rule

encouraged the exploration of alternatives and the second

rule encouraged the exploitation of the most promising

solutions. It generated the disassembly completed graph

dynamically and implicitly during the search process of the

ants, the assembly sequence was obtained by reversing the

search path of the ants. Several other heuristic algorithms,

such as artificial immune systems [21] and particle swarm

optimization algorithm [22], also have been used to solve

the assembly sequence planning problem.

The advantages and disadvantages of the above three

kinds of method for ASP are listed in Table 1.

In this paper, a memetic algor ithm is used to solve the

assembly sequence planning problem. Memetic algorithm

(MA), firstly proposed by Moscato [23], is a kind of

stochastic global search heuristics in which evolutionary

algorithms-based ap proaches are combined with local

search techniques to improve the quality of the solutions

created by evolution. Compared with GA, MA is able to

make a better exploration of the neighborhood. MA has

1176 Int J Adv Manuf Technol (2010) 49:1175–1184

Author's personal copy

剩余10页未读，继续阅读

weixin_38599518

粉丝: 7
资源: 882

改进的 memetic 算法在产品装配序列规划中的应用

Memetic Algorithm.zip_MA算法.zip_memetic algorithm_启发式算法_文化基因算法_遗传

文化基因算法简介（Memetic Algorithm，MA）

Memetic Algorithm(文化基因算法)

memetic algorithm代码

memetic q-learning

python实现Memetic算法程序设计及应用

使用一种具体的局部搜索算法设计一个memetic算法框架，画出详细的流程图，给出步骤说明以及使用了哪种局部搜索算法。

设计一个memetic框架，并具体使用一种搜索策略，写出实现代码，并写出流程

python实现Memetic算法程序设计及应用并生成3D图像

设计一个memetic

最新资源