• No se han encontrado resultados

1. Preliminares

1.4 Antecedentes

As outlined in Section 3.1,the decision-making in a DSS is very complex when any phase of decision-making is non-structured, which causes the final decision choosing process uncertain. This is quite common in decision making in water resource management problems. Decision support can be carried out either simply by searching results from likely decision scenarios to answer the “what if “ question related to the problem, or by developing a large and complex Expert System. The Decision Support subsystem in DSSFCMR was developed to provide decision support through data searching of the simulated flood scenarios stored in the database. The data in DSSFCMR is the collection of decision variables such as the peak discharges, the water levels and so on. Some information on database and data search support are introduced in this section.

To define and understand exactly what is meant by the database, first the data should be described. Data are the facts about an object or a concept. Data can be a person or an organization (e.g. the decision maker for flood warning or an user of a DSS), a place (e.g. catchment, house), an event (e.g. a flood event), an action (e.g. flood forecasting and warning) and many others, or they can be any combination of above facts. A single fact could be considered as an element of data. Information is data that are organized, refined and presented in a form that can be used to aid the decision-making process or alos perform other activities such as identifying the flood-inundated houses and drawing the flooded area. Often data

considered as a unit of information. The difference between the two definitions of data and information can become uncertain and unclear.

A database is a shared collection of logically related data and their description. The data in the database is designed to meet information requirements of an organization or a project. A database management system (DBMS) is a system that allows the users to define, create, and maintain the database, and offers controlled access to this database. The advantages of database management systems include control of data redundancy; data consistency; extracting more information from the same amount of data; sharing of data; improved data integrity; improved data security; enforcement of data standards; improved data accessibility and responsiveness; increased productivity and more. However, there are also some disadvantages such as complexity, cost of developing and maintaining DBMSs, additional hardware costs, cost of maintenance and so on.

DBMS in general consists of several databases (each having several tables), which are represented by data models. The data model is an integrated collection of concepts of expressing data, relationships between data and constraints on the data in an organization. The relational data model, which is widely used in MS ACCESS, is based on the concept of mathematical relations. In a relational data model, data and relationships are represented as database tables, and each database table has a number of columns with unique names. These can be considered as the parameters of the database. A primary key, which consists of one or several parameters, uniquely identifies a record in a database table, which can help to locate the data. The relational data model is very popular and is most widely used in Personal Computers (PCs). It has become the de-facto standard for the design of both large and small databases. The relational database model was used in DSSFCMR in DBMS.

The decision making support in this research project was developed through information search, and such information includes the peak discharge and flood water levels along the flood inundated area, stored in the database. Data search was developed based on the Structured Query Language or SQL as it is commonly called (Connolly et al., 1999). The SQL language is a particular language that has emerged from the development of the relational model. It is a non-procedural language and is relatively easy to learn. SQL is essentially free-format which means that the statements do not have to be typed at particular locations on the screen. SQL has two major components:

• A Data Definition Language (DDL) for defining the database structure.

• A Data Manipulation Language (DML) for retrieving and updating data.

The Data Definition Language (DDL) was used in this research project to define the database structure (e.g. create a database table). The Data Manipulation Language (DML) was used to develop the Windows interface to help the user to locate the primary key for the decision support. Therefore, the basic idea of DML is introduced here:

The SQL DML statements have four types of command functions as described below:

• SELECT: To find data in the database

• INSERT: To add data to a table in the database

• UPDATE: To modify the data in a table in the database

• DELETE: To delete data from a table in the database

For example, a SQL statement is required to be written to find all scenarios with minimum of 1.05 for Alpha, a minimum value of 0.8 for m, a minimum of 1 mm/h for continuing loss and a maximum value of 20 cumecs for base flow in the URBS model, with 3 hours of forecast rainfall with the forecast greater than 0 mm and the forecast in the first hour of 20% of the total, and the decision made at 01 January 1983. The data of all scenarios are stored in database table RtFw1 for Real- Time Flood Warning. The SQL statement for this example is given below:

Select * from RtFw1 where Alpha >= 1.05 and m >= 0.8 and ContLoss >= 1 and BaseFlow <= 20 and ForecastStormPeriod = 3 and ForecastStormQty01 >= 0 and ForecastStorm01P1 >=20 and Year = 1983 and Month = 1 and Day = 1

Through the developed user-friendly Windows interface in DSSFCMR, the decision maker can find the likely decision scenarios from many scenarios produced by the models (in the Modelbase), using the database search technology and SQL statements. Using these likely decision scenarios coupled with the decision maker’s own experience, the decision maker can finalize the decision for flood warning. The system does not demand the decision maker to require the database and SQL knowledge to run the DSSFCMR for decision making.

The design and development of the Decision Support subsystem will be detailed in Chapters 4 and 5.

3.6 Summary

The relevant concepts, theory and models in DSS design and development, the hydrological model, the hydraulic model, the spatial and graphical data display and analysis (SGDDA), and the decision support are introduced in this chapter. The purpose of the chapter was not to introduce all theory and practices, but to describe the relevant and sufficient theory and knowledge, which will assist further design and development, which will be described in Chapters 4 and 5.

In the sections of DSS design and development, the hydrological model, the hydraulic model and the decision search support, only the selected method or model is detailed. These particular methods or models have been selected after careful examination of their reputation, easy integration, and the support of supplementary software tools for the potential user. However, there is no doubt that DSSFCMR can co-operate with other models if they are found to be more suitable. For example, the RORB (or WBNM) hydrological model can be used instead of the URBS model, but this requires certain integration works for linking RORB (or WBNM) within DSSFCMR, which are different to linking URBS (but on similar principles).

The SGDDA was a significant development in integration of outputs of hydrological and hydraulic models with GIS software tools, and in performance of spatial analysis for flood scenario decisions in DSSFCMR, which requires handling of DXF files, creation of shapefiles and use of MapObjects software.

The theory and knowledge presented in this chapter can also be used and extended in future for further system design and development of DSSFCMR (and other DSSs).