Real-Time Decision-Making with Partial Information for Construction Management

Real-time decision-making with partial information is commonplace in the daily tasks of a construction manager/engineer. However, traditional decision support systems (DSSs) do not support partial information inference. Data pre-processing was adopted to solve the data-incompleteness problem. Unfortunately, such approach may be biased. This paper presents a newly developed neuro-fuzzy system, named Variable-attribute Fuzzy Adaptive Logic Control Network (VaFALCON), for decision-making under partial information environment with the untreated original incomplete data. Two case studies with 91.7% and 83.3% attribute information were conducted to test the proposed VaFALCON system. It is found that the average accuracy recovery ratio (AARR) is between 90.57% and 95.68% for testing data with 91.7% partial information, and is between 86.03% and 93.67% for testing data with 83.3% partial information.