Abstract—To create a solution for a specific problem in machine learning, the solution is constructed from the data or by use a search method. Genetic algorithms are a model of machine learning that can be used to find nearest optimal solution. While the great advantage of genetic algorithms is the fact that they find a solution through evolution, this is also the biggest disadvantage. Evolution is inductive, in nature life does not evolve towards a good solution but it evolves away from bad circumstances. This can cause a species to evolve into an evolutionary dead end. In order to reduce the effect of this disadvantage we propose a new a learning tool (criteria) which can be included into the genetic algorithms generations to compare the previous population and the current population and then decide whether is effective to continue with the previous population or the current population, the proposed learning tool is called as Keeping Efficient Population (KEP). We applied a GA based on KEP to the production line layout problem, as a result KEP keep the evaluation direction increases and stops any deviation in the evaluation. Keywords—Genetic algorithms, Layout problem, Machine learning, Production system. I. INTRODUCTION O create a solution to a specific problem in machine learning, the solution is constructed from the data or by use a search method to find nearest optimal solution. Genetic algorithm (GA) is stochastic search which is often used in machine learning application. An effective GA representation and meaningful fitness evaluation is the keys of the success in GA applications. The effectiveness of GA comes from their simplicity as robust search algorithm as well as from their power to discover good solutions rapidly for difficult high-dimensional problems. GA is useful and efficient in the following cases: Ԙ When the search space is large, complex or poorly understood. ԙ When the domain knowledge is scarce or expert knowledge is difficult to encode to narrow the search space. Ԛ If no mathematical analysis is available. ԛ When the traditional search methods fail. The advantage of the GA approach is that it can handle Manuscript received November 30, 2006. A. J. Author is with Intelligent Manufacturing Systems Laboratory, Gifu University. 1-1 Yanagido, Gifu Shi, 501-1193, Japan. (e-mail: k3812203@edu.gifu-u.ac.jp) Y. H. Author, is with Intelligent Manufacturing Systems Laboratory, Gifu University. 1-1 Yanagido, Gifu Shi, 501-1193, Japan. (corresponding author, phone: 81-293-2550; fax:81-293-2550;.e-mail: yam-h@gifu-u.ac.jp). R. R. Author is with Intelligent Manufacturing Systems Laboratory, Gifu University. 1-1 Yanagido, Gifu Shi, 501-1193, Japan. (e-mail: k3812205@edu.gifu-u.ac.jp) arbitrary kinds of constraints and objectives; all such things can be handled as weighted components of the fitness function, making it easy to adapt the GA scheduler to the particular requirements of a very wide range of possible overall objectives. GA is one of the wide applicability of knowledge-learn problem-solving tools that can provide an approach [1], [2]. While the great advantage of GA is the fact that they find a solution through evolution, this is also the biggest disadvantage. Evolution is inductive; in nature life does not evolve towards a good solution but it evolves away from bad circumstances. This can cause a species to evolve into an evolutionary dead end. In order to reduce the effect of this disadvantage a learning tool is add to the GA to decide whether is effective to continue with the generated population or the current population, this learning tool is called hereafter as keeping efficient population (KEP). II. GENETIC ALGORITHMS The GA paradigm has been proposed to solve a wide range of problems [2][3][4]. GA has been successfully applied to optimization problems in diverse fields and it is differs from other search techniques which depend on natural genetic evaluation process. GA starts with an initial set of solutions selected randomly called population. A suitable encoding for each solution in the population is used to allow computation of the fitness. The solution set in the population, called as chromosome or individual, represents a solution to the optimization problem. Each individual contains a number of genes. The individuals in the initial population are evaluated to measure its fitness. To create the next population, new individuals are formed by either merging two individuals from the current population using a crossover operator or modifying an individual solution using mutation operator. Based on the individuals’ fitness, the individuals to be included in the next population are then probabilistically selected from the set of individuals in current population. The iteration, called a generation is continued until the fitness reaches its maximum value, with the hope that strong parent will create a fitter generation of the children. The best overall solution becomes the candidate solution to the problem. To create the next generation GA based on three operations: Selection, crossover and mutation. A. Basic Genetic Algorithm The outline of the basic GA is described below. 1. [Start] Generate random population of n individuals (suitable solutions for the problem). Machine Learning in Production Systems Design Using Genetic Algorithms Abu Qudeiri Jaber, Yamamoto Hidehiko and Rizauddin Ramli T International Journal of Computational Intelligence 4;1 © www.waset.org Winter 2008 72