An adopted neural network model for estimating productivity performance of ready-mixed concrete (RMC) placement for Metro Manila buildings

• Main Author: Alonzo, Joel U.
• Format: text
• Language: English
• Published: Animo Repository 1999
• Subjects: Ready-mixed concrete; Concrete construction; Construction Engineering and Management
• Online Access: https://animorepository.dlsu.edu.ph/etd_masteral/6831
• Institution: De La Salle University

A neural · network technique was used in this study to adopt a model in predicting ready-mixed concrete placement productivity specifically by concrete pumps and by crane arid buckets, and to identify the key factors affecting it. Data from 27 surveyed building projects in Metro Manila have been used. The study used a structured survey questionnaire as the primary data-gathering instrument to gather project information and input from experienced project respondents, while ocular inspections, personal interviews as secondary instruments. The structure of the adopted neural network model was developed in a spreadsheet format that is customary to many construction practitioners, and to provide a more transparent and simplified representation of this technique. A user-friendly interface was developed using spreadsheet macros to simplify user input and automate estimate prediction. For practicality the model could be changed to suit the experience, and the environment that enables the user to become more suited to the model. It also enables the build up of experience and incorporates new encounters into the model. In order for the model to recognize new patterns, the model needs to be retrained with new patterns along with previously known patterns. If only new patterns were provided for retraining, then the old patterns may be forgotten. Instead of using the back-propagation algorithm in finding the optimum weights or training the adopted neural network model, simplex optimization had been used. The model is able to give a good RMC placement productivity estimate and altogether 18 key factors were identified covering the areas/categories of the project manager, design and planning, skilled labor, material supply, work environment, project controls, building specifications, and crew size. The model can be used to evaluate the various factors and thus RMC placement productivity can be effectively managed, monitored and deployed.