Presenting a super-heuristic genetic algorithm for investment in project resource
محورهای موضوعی : production operation management
1 - Geisha St. 13, Plate 40.
کلید واژه: Genetic Algorithm, Project Scheduling, Resource Limitation,
چکیده مقاله :
Developing a suitable plan and optimal use of available facilities are considered important factors in today's competitive world. The aim of this research is to provide an innovative genetic algorithm for the problem of investment in project resources. In terms of the purpose, this research is an applied and, in terms of data collection, it is of a mathematical analytical type. According to the positive experiences of using genetic algorithm to solve the problems of the specification in limited resources, this research aims to create two genetic algorithms for a type of allocation problem called investment problem in resources. Genetic algorithm designed was tested on the problems investigated by Mohring representing that the above problems are not complicated enough, because genetic algorithm has obtained optimal solution for the problems rapidly. So, more problems were generated by Progen software through more tests, and, in general, more than 15,000 problems tested by genetic algorithm. Then, by making changes in the above algorithm and using Akpan method and modifying this method, genetic algorithm has been improved. The method developed has also been compared with the previous method during the tests. After setting the parameters on 20 activity problems, the tests were conducted on 10 and 14 activity problems. It represented that new algorithm works more efficiently on these problems. On 30 activity problems in Dergzel and Kims, new and previous genetic algorithms were compared by using multivariate variance analysis and Duncan's test indicating a significant improvement in the answers.
Developing a suitable plan and optimal use of available facilities are considered important factors in today's competitive world. The aim of this research is to provide an innovative genetic algorithm for the problem of investment in project resources. In terms of the purpose, this research is an applied and, in terms of data collection, it is of a mathematical analytical type. According to the positive experiences of using genetic algorithm to solve the problems of the specification in limited resources, this research aims to create two genetic algorithms for a type of allocation problem called investment problem in resources. Genetic algorithm designed was tested on the problems investigated by Mohring representing that the above problems are not complicated enough, because genetic algorithm has obtained optimal solution for the problems rapidly. So, more problems were generated by Progen software through more tests, and, in general, more than 15,000 problems tested by genetic algorithm. Then, by making changes in the above algorithm and using Akpan method and modifying this method, genetic algorithm has been improved. The method developed has also been compared with the previous method during the tests. After setting the parameters on 20 activity problems, the tests were conducted on 10 and 14 activity problems. It represented that new algorithm works more efficiently on these problems. On 30 activity problems in Dergzel and Kims, new and previous genetic algorithms were compared by using multivariate variance analysis and Duncan's test indicating a significant improvement in the answers.
