Expert systems

Expert systems are processes that mimic the way in which humans diagnose problems and dispense advice (Mumford and Norton, 1989, 1993). They are usually computer based although for practical implementation it is often more appropriate to transfer the expert system to a paper or manual format. The steps involved in developing an expert system include problem structuring, knowledge acquisition, knowledge engineering and encoding, verification, validation and testing (Mumford and Norton, 1993). The basis of expert systems is a set of linked logical relationships of rules with the structure 'If some condition^) then some conclusion(s) else other conclusion(s)' (Holt and Cheke, 1997).

A very simple programme could have one rule:

If pest population is above threshold

(quality value) then spray (choice)

else do not spray (choice)

A number of expert systems have been developed for pest management use (Jones et al., 1990; Compton et al., 1992; Knight et al., 1992; Tang et al., 1994).

0 0

Post a comment