روش اطلاعاتی برای سیستم برنامه ریزی تولید برای محصولات پیش ساخته بتن سفارشی

Bespoke precast concrete products are widely used components of construction projects. These products implement the offsite prefabrication technology that offers a unique opportunity for innovation and cost savings for construction projects. However, the production process from design to manufacturing contains uncertainties due to external factors: multi-disciplinary design, progress on construction site. The typical workload on bespoke precast factories is a complex combination of uniquely and identically designed products, which have various delivery dates and requirement of costly purpose-built moulds. In this context, this research is aimed to improve the efficiency of the process by addressing the production planning because it has a significant impact to the success of the production programme. An innovative planning system and its prototype called ‘Artificial Intelligence Planner’ (AIP) are developed. AIP is capable of two functionalities. The first is a data integration system that encourages the automation in the planning process. The other is a decision support system for planners to improve the efficiency of the production plans. These functionalities reinforce each other to deliver optimum benefits to precast manufacturers. AIP have employed artificial intelligence technologies: neural network and genetic algorithm to enhance data analyses for being a decision support for production planning. The outcomes of the research include shortened customer lead-time, in-house repository of production knowledge, and achievement of the optimum factory's resource utilisation.

The precast industry is a major supplier of offsite-prefabricated components to the construction industry. The construction of a building can be regarded as an assembly of hundreds of different designs and delivery dates of ‘bespoke’ precast concrete units. This demand creates the difficulty in the bespoke precast production. ‘Bespoke’ precast production system is in ‘make-to-order’ or ‘engineer-to-order’ style. Bespoke precast production has a major distinction from ordinary ‘mass production’ that every time the process is started from new product design. The complexity of bespoke precast production is based on this ground. Since the production is less uniform, the ‘learning curve’ is hard to establish and the automation is hardly implemented to assist the process. The optimum resources utilisations are serious issues of precast manufacturers. The production planning of this product kind requires sophisticated management and it is a key of the success of the delivery program, customer lead-time competitiveness, and the effective utilisation of purposed-built precast mould. Despite these, the current practice of the production planning is very plain and effortful. The earliest due date rule (EDD) is traditionally being used by the bespoke precast industry as a scheduling method. It is a simplified method regardless of resources considerations [1]. The aim of this research is to develop a new (semi-automatic) planning system to manage bespoke precast production called the ‘Artificial Intelligence Planner’ (AIP). AIP is designed to assist the production planning process. It functions as data integration and decision support system that supports the automation in the planning process and increases the efficiency of production plans. AIP employs a kind of technologies called ‘artificial intelligence’ (AI). AI typically is composed of sophisticated algorithms, which have the ability of analysing and processing data similar to human's intelligence. AIP powered with these AI technologies can be implemented by planners as a decision support system and help reduce human involvement (consequently, increase automation) in the planning process. In fact, the AI technologies have been introduced to the construction and manufacturing industries for many decades. Its successful applications and usefulness have been presented in many research studies and are increasing by the time. Two AI used in AIP are ‘neural networks’ (NN) that is applied to estimate the manufacturing time required for any unique product design and ‘genetic algorithm’ (GA) that is used in the scheduling optimisation to search for optimal results. In addition, this AIP system conforms to IFC (Industry Foundation Classes), which is the ongoing development aimed to integrate all processes of the whole life cycle of construction projects. The application of AIP can help include a part of precast concrete usages and managements [21].

The bespoke precast production is currently dealing with a high complexity due to various product designs and assorted delivery dates. The traditional production planning tasks are very manual and simple by applying EDD to arrange their production sequence. This research initiated an innovation production planning system called AIP that adopted artificial intelligence technologies to alleviate this complexity and improve the efficiency in the bespoke precast concrete production planning. The result from the experimentation has shown that the reorganisation of CAD drawing elements together with rules-based objective recognition can automate the material quantity take-off task. NN with MR techniques were implemented together to estimate processing times of manufacturing routines. Also, the GA-based optimisation on the proposed flowshop scheduling model can provide statistically better schedules than the traditional ones from EDD around 25% (total flowtime reduction). In addition, these planning tasks are integrated together to establish automation in the process. The applicability of the AIP system was also evaluated through a real case study using input from another company. The outcomes of the system include shortened customer lead-time, in-house repository of production knowledge, and optimum factory's resource utilisation. This improvement can benefit to both precast concrete and construction industries. Finally, the concept of AIP could be applied to other kinds of offsite prefabrication products, which also require the make-to-order production system so that their production could be linked to the information from project team and construction site.