Designing the layout of single- and multiple-rows flexible manufacturing system by genetic algorithms
Introduction
Layout of flexible manufacturing system (FMS) involves distributing different resources in a given FMS and achieving maximum efficiency of the services offered. With this in mind FMS are designed to optimize production flow from the first stages as raw material to finished product. The layout has an important impact on the production time and cost, especially in the case of large FMS [1]. It was estimated that 20–50% of the manufacturing costs are due to handling of work pieces; by a good arrangement of devices it is possible to reduce the manufacturing costs for 10–30% [2]. Some other authors report even higher percentage of material handling based costs, for example Chiang and Kouvelis report that 30–70% of total manufacturing costs may be attributed to materials handling and layout [3]. Therefore, already in an early stage of designing of the FMS it is necessary to have an idea of the layout of the devices.
Usually the selected fitness function is the minimum total costs of handling of work pieces. In general, those costs are the sum of the transport costs (these are proportional to the intensity of the flow and distances) and other costs.
Section 1 of the paper presents the problem and the aim of designing the FMS. Section 2 introduces the FMS and its specific properties with respect to transport and design. Section 3 makes survey of researches in the area of designing the FMS. Section 4 briefs the reader on the genetic algorithms (GA) method used in our work. Section 6 gives detailed information about the method itself of searching for solution by GA and the evolutionary and genetic operators used. Section 7 summarizes the results obtained by the model. The discussion and the concluding findings follow.
Section snippets
Characteristic of layout of FMS
The manner of arranging of working devices largely depends on the type of production [4]. The FMS have some specific properties decisively influencing the designing and the construction of the system.
The FMS started to be used due to changed conditions on the market. The following two trends, which have greatest influence on the change of manufacturing system, can be noticed [5]:
- •
Influence of customer's special desires and requirements, i.e., greater complexity and versions of products;
- •
Increased
Survey of facility layout methods
Designing a FMS has to do with arranging unequally large devices. Therefore, only the methods for arranging differently large devices can be used. Generally, unequal-area layout problems are more difficult to solve than equal-area layout problems, primarily because unequal-area layout problems introduce additional constraints into the problem formulation [7].
The problem of arranging of devices is one of so-called NP problems. NP-hard problems are unsolvable in polynomial time [4]. Accurate
Genetic algorithms
GAs are a new approach to solving complex problems such as determination of facility layout. They can be defined as meta-heuristics based on the evolutionary process of natural systems [13]. GAs became known through the work of John Holland in the 1960s.
The GAs contain the elements of the methods of blind searching for the solution and of directed and stochastic searching and thus give compromise between the utilization and searching for solution. At the beginning, they search in the entire
Model of designing the FMS
As already mentioned the most suitable form of arrangement of the devices in the FMS for serving by AGV is in a single or in multiple rows. Prior to solving the problem we set ourselves the limitations that will be taken into account. These limitations are:
- •
all machines are of rectangular shapes;
- •
all machines are operated in the center of that space;
- •
the available surface for FMS is rectangular in shape;
- •
the available surface for FMS is limited along width;
- •
all machines in the row are oriented as
Coding of organisms, fitness function and genetic operations
The main steps in solving of the problem with GA used are represented in Fig. 3.
In the first step of the GA the initial population is created at random. Only correct organisms representing feasible solution are created. This initial generation enters the evolutionary loop of the GA. After evaluation of the population the selection of organisms with the tournament method is made. In the next step the operations of reproduction and crossover with probability pr and pc, follow respectively. The
Results and discussion
For testing of our model the test example forming of FMS with 14 devices was used. It can be calculated that more than 8.7 × 1010 different forms of FMS are possible. First, we collected all necessary data for designing of the FMS. On the basis of that data two types of FMS were designed: a single-row FMS (a is not limited), and a multiple-rows FMS (a = 12.5 m). In the preliminary runs of GA system a combination of evolutionary parameters with acceptable probability of success was searched.
The
Conclusions
By means of the presented model we can find the optimum layout of the devices in the FMS. The model searches for the optimum layout in rows and finds itself the optimum numbers of rows. The layout can be either the layout in single row or multiple rows. The model does not limit itself to one solution only, but it can propose several equally good solutions which can differ very much. From the solutions reached, having similar values of the fitness functions, the rules for designing our FMS by
References (18)
- et al.
A hybrid heuristics for the facility layout problem
Comput. Operat. Res.
(1991) - et al.
The facility layout problem
Eur. J. Operat. Res.
(1987) - A. Hamiani, G. Popescu, CONSITE: a knowledge-based expert system for site layout, Computing in Civil Engineering:...
- et al.
Facilities Planning
(1996) - et al.
Improved Tabu search heuristics for heuristics for solving facility layout problems
Int. J. Prod. Res.
(1996) Intelligent Manufacturing Systems
(1990)Flexible Manufacturing Systems
(2001)- et al.
Machine layout problem in flexible manufacturing systems
Operat. Res.
(1988) - L. Heng, P.E.D. Love, Genetic Search for Solving Construction Site-Level Unequal-Area Facility Layout Problems,...
Cited by (67)
Analysis of drivers for solving facility layout problems: Literature review
2021, Journal of Industrial Information IntegrationDigital twin driven factory design
2020, Digital Twin Driven Smart DesignSolving an extended multi-row facility layout problem with fuzzy clearances using GA
2017, Applied Soft ComputingA dynamic multi-objective approach for the reconfigurable multi-facility layout problem
2017, Journal of Manufacturing SystemsSingle row facility layout using multi-start simulated annealing
2017, Computers and Industrial EngineeringSolving dynamic double row layout problem via combining simulated annealing and mathematical programming
2015, Applied Soft Computing JournalCitation Excerpt :MRLP involves multiple rows and is to place facilities on those rows and each pair of facilities is separated by their minimum clearances. This problem is usually formulated as a quadratic assignment problem (QAP) [16] and some meta-heuristics have been proposed to solve it [17,18]. Both SRLP and MRLP are combinatorial due to the sequence of facilities being considered on one row or multiple rows.