Introduction using selection, reproduction, crossover and mutation. Uses


Numerical models are used when analysing coastal water
process to model the flow and quantify the problem. Yet current models are inadequately user-friendly, relying human expertise,
causing the needed information to solve the problem to not be translated into
the model. AI technology can now imitate that human expertise when problem solving.
Hence the review of
current development and progress of integrating AI into water quality models.

need for integration with AI                                                                                                                 

Current parameters do not fully reflect knowledge of
real observations, mathematical description of water movement chemical process
etc, when manipulating the model. These impacts can lead to complete failure of
model.                                                                                                It is common practice to use existing
knowledge and to only change one or two parameters during model manipulation to
simply the problem without losing direction. However, AI techniques are not
only able to implement existing knowledge but won’t get lost in the
manipulation direction.                                                           
can be noted that fifth generation models have incorporated AI technology.

Integration with artificial intelligence


an increasing demand for AI systems to be integrated and for models to be
improved, a categorization of various AI techniques has been produced:


1.     Knowledge-based systems (KBSs)-Using a
Symbolic and logical reasoning algorithm, it mimics and automates the decision-making
and reasoning process of human experts. It enables the selection and manipulation
of various numerical models on hydrodynamics/water quality

2.     Genetic algorithms (GAs)-Evolutionary
algorithm that using selection, reproduction, crossover and mutation. Uses
computational model of natural evolutionary process. Optimized calibration of the
parameters of numerical models on hydrodynamics and water quality.           

3.     Artificial neural networks (ANNs)-Are
based on present understanding of the brain and nervous system. Its application
comes in by determining physical/biological relationships that are not fully
understood as well as optimized calibration.

4.     Fuzzy inference systems-Uses the fuzzy
set theory to map elements of the fuzzy set to universe of member. This can be
made use of by quantifying undefined and to determine their confidence factors.                                                                                                                        







Future direction

                       The versatility
of water quality models can be increased by combing two or more of the explored
options to create a hybrid combination. More research and developments are
being made in AI technology which will enhance its use and technology in water
quality modelling, along with ease of use.



existing water quality models not being user friendly needing specialist
training. The advancement in AI technology have been provided a way to bridge
that gap. Attempts have been made to integrate systems such as KBS, GA, ANN, and fuzzy interference system into numerical models. With the
increasing capabilities of AI technology, the future developments of numerical
modelling looks promising.