Scholarlink

<p>In this study, a new method based on concurrency engineering in big construction projects is presented. The proposed method is to use a training model based on the perceptron neural network. In this method, which consists of two phases of training and testing, a database of the project is formed based on the time duration information related to the tasks and the sources of performing the tasks. Then in the next step, the tasks are weighted and based on the perceptron network, the weights are updated until the minimum execution time is reached. Next, the task database is formed based on the timetable and based on the most to the least weight in the form of training data. Finally, a concurrency engineering pattern is formed on this table. The proposed method is a new method for forming concurrency engineering in large projects. Time reduction percentage, concurrency error percentage and cost reduction percentage are considered as evaluation indicators that the results have improved by 39%, 22% and 27%, respectively. However, with comparison in the proposed perceptron neural model versus LSTM, the proposed perceptron neural model shows better performance than LSTM.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using combined method based on ANN-concurrent to time-cost estimation in projects

  • 期刊论文

  • 发表日期：

    2026年5月13日

查看全文
  • Kaveh Ostad-Ali-Askari,
  • Hamid Jafarinia,
  • Peiman Kianmehr

摘要

In this study, a new method based on concurrency engineering in big construction projects is presented. The proposed method is to use a training model based on the perceptron neural network. In this method, which consists of two phases of training and testing, a database of the project is formed based on the time duration information related to the tasks and the sources of performing the tasks. Then in the next step, the tasks are weighted and based on the perceptron network, the weights are updated until the minimum execution time is reached. Next, the task database is formed based on the timetable and based on the most to the least weight in the form of training data. Finally, a concurrency engineering pattern is formed on this table. The proposed method is a new method for forming concurrency engineering in large projects. Time reduction percentage, concurrency error percentage and cost reduction percentage are considered as evaluation indicators that the results have improved by 39%, 22% and 27%, respectively. However, with comparison in the proposed perceptron neural model versus LSTM, the proposed perceptron neural model shows better performance than LSTM.