【正文】 um problem of this text aim convection solve a method to carry on more analytical。t very ideal in physically the application. The dynamic plan (dynamic programming) the algorithm is one monly used method that solves the multi stages decisionmaking process optimization question。目前，對(duì)于裝配線平衡問(wèn)題的研究主要分為以下 3 個(gè)方面：給定裝配線的節(jié)拍，求最小工作站數(shù)，通常在裝配線的設(shè)計(jì)與安裝階段進(jìn)行；給定裝配線的最小工作站數(shù)，使裝配線的節(jié)拍最小，對(duì)已存在的生產(chǎn)線進(jìn)行調(diào)整優(yōu)化；在裝配線的工作站數(shù)和節(jié)拍得到優(yōu)化確定的條件下，均稀裝配線上工作站的負(fù)荷，給工作人員一種公平感。我們必須另 尋其他方法。 裝配線平衡問(wèn)題經(jīng)歷了一個(gè)由最優(yōu)化算法到啟發(fā)式算法，直到現(xiàn)代人 工智能方法和工業(yè)工程方法的過(guò)程。王云鳳等人運(yùn)用混合遺傳算法求解裝配線平衡問(wèn)題，該算法將啟發(fā)式規(guī)則、遺傳算法、模擬退火算法綜合在一起，既利用了啟發(fā)式規(guī)則基于經(jīng)驗(yàn)的優(yōu)勢(shì)，可以提高算法的速 9 度，又可使遺傳算法和模擬退火算法優(yōu)勢(shì)互補(bǔ)，有效地解決了裝配線平衡問(wèn)題。由于運(yùn)用線性規(guī)劃模型方法求解運(yùn)算量很大，只能求解作業(yè)元素個(gè)數(shù)較少的線平衡問(wèn)題，與實(shí)際情況相差甚大，故此法在實(shí)際應(yīng)用中所取得的效果不是很理想 。除了造成無(wú)謂的工時(shí)損失外，還造成大量的工序堆積，嚴(yán)重時(shí)會(huì)造成生產(chǎn)線的中止。 The minimum work station given to settle assembly line number, make the rhythm of assembly line minimum, to already exist of the production line carry on adjust excellent turn。 it is very difficult and skillful. The dynamic plan realization is not certainly plex, is suitable for many questions, when solution general question is one of we first choice algorithms. But, the dynamic 4 plan mathematical model establishment is not an easy matter, most difficult also most importantly condition expression. The dynamic plan condition expressed the description subquestion must satisfy the most superior child constitutive property, otherwise unable to establish the correct dynamic plan model. Solves the problem when the application dynamic plan method, should first estimate the question the time, the space, if the question has the dimension barrier, then the dynamic plan condition expressed very difficult to satisfy the paratively largescale question the spatial request. We must seek the alternative means in addition. The dynamic plan law utilizes to when solving the ALB problem, its condition expression with various workstations work element process period related, often when determination dynamic upper boundary must affect a smaller condition to the system to carry on the deletion, thus achieved reduces the operand the goal. The heuristic method by its simple, easy to understand, was fast and has satisfied won the multitudinous administrative personnel39。由于裝配 線的平衡總是離散型組合優(yōu)化問(wèn)題，尋求最優(yōu)解比較困難，故采取合適的方法解決裝配線平衡問(wèn)題引起了各企業(yè)界和學(xué)術(shù)界的廣泛關(guān)注。動(dòng)態(tài)規(guī)劃法運(yùn)用于解決 ALB 問(wèn)題時(shí)，其狀態(tài)的表示與各工作站作業(yè)元素的加工時(shí)間有關(guān)，往往在確定動(dòng)態(tài)上界時(shí)需對(duì)系統(tǒng)影響較小的狀態(tài)進(jìn)行刪除，從而達(dá)到降低運(yùn)算量的目的。本文針對(duì)線性規(guī)劃法、動(dòng)態(tài)規(guī)劃算法、遺傳算法和工業(yè)工程方法這四種求解裝配線平衡問(wèn)題的典型方法分別作了簡(jiǎn)要分析，指出了各種算法的優(yōu)缺點(diǎn)及應(yīng)用情況。遺傳算法在實(shí)際的應(yīng)用中往往出現(xiàn)早熟收斂和收斂性能差等缺點(diǎn)，現(xiàn)今的一些改進(jìn)方法大都是針對(duì)基因操作、種群的宏觀操作、基于知識(shí)的操作和并行化 GA進(jìn)行。之后一些學(xué)者將此算法加以改進(jìn)，但運(yùn)算量也比較大。制造業(yè)的生產(chǎn)多半是在進(jìn)行細(xì)分化之后的多工序、流水化、連續(xù)作業(yè)生產(chǎn)線，此時(shí)由于分工作業(yè)，各工序的作業(yè)時(shí) 間在理論上、實(shí)際操作上都不能完全相同，這就勢(shì)必存在工序間作業(yè)負(fù)荷不均的現(xiàn)象。s production efficiency of the car assembly line. The assembly line equilibrium problem is the longlost type of a type of typical model the bination is excellent to turn a problem, particularly is for random, many the assembly line equilibrium problem of target, solve to the satisfaction seldom more on a certain degree. Mainly is divided into the following 3 aspects to the research of assembly line equilibrium problem currently: Give the rhythm of the assembly line certainly beg minimum work station number, usually in the assembly line of design and install the stage carry on。s trust and the wele, and is widely applied to each domain. Its production mainly i