![]() |
|||
![]()
|
![]() |
![]() |
![]() |
3. HISTORICAL ACCOUNTKnowledge-based expert system technology has been applied to a variety of agricultural problems since the early 1980s. The following paragraphs present how expert systems were considered in agriculture in the 1980s. The papers have been selected to represent different applications and to be easily obtained by interested readers. The expert system applied to the problems of diagnosing soybean diseases (Michalski et al., 1983) is one of the earliest expert systems developed in agriculture. A unique feature of the system is that it uses two types of decision rules: (1) the rules representing expert diagnostic knowledge, and (2) the rules obtained through inductive learning from several hundred cases of disease. Experimental testing of the system has indicated a high level of correctness of the system's advice (in an experiment involving a few hundred cases , approximately 98% of the diagnosis were correct). POMME (Roach et al,, 1985) is an expert system for apple orchid management. POMME advises growers about when and what to spray on their apples to avoid infestations. The system also provides advice regarding treatment of winter injuries, drought control, and multiple insect problems. COMAX (Lemon, 1986) is a crop management expert system for cotton that can predict crop growth and yield in response to external weather variables, physical soil parameters, soil fertility, and pest damage. The expert system is integrated with a computer model, Gossym, that simulates the growth of the cotton plant. This was the first integration of an expert system with a simulation model for daily use in farm management. In 1987, expert system technology was identified as an appropriate technology to speed up agricultural desert development in Egypt (Rafea and El-Beltagy, 1987). The paper discussed the importance of applying expert systems in agricultural desert development in Egypt and suggested an integrated structure of an R&D unit to develop and maintain an efficient use of these systems. CALEX system (Plant, 1989) was developed for agriculture management. It is domain independent and can be used with any commodity. CALEX consists of three separate modules: an executive, a scheduler, and an expert system shell. The executive serves as the primary interface to the user, to models, and to the disk. The scheduler generates a sequence of management activities by repeatedly activating the expert system. The expert system makes the actual management decisions. Initial development of the system has focused on the development of a package of modules for California cotton and another package for peaches. In the 1990s, several expert systems have also been developed. A sampling of these expert systems has also been selected as examples. The selected examples demonstrate trends in the 1990s and their accessibility by the reader is also considered. An agroforestry expert system (UNU-AES) was designed to support land-use officials, research scientists, farmers, and other individuals interested in maximizing benefits gained from applying agroforestry management techniques in developing countries (Warkentin et al., 1990). UNU-AES is a first attempt to apply expert systems technology to agroforestry. This system addresses the option for alley cropping, a promising agroforestry technology that has potential applicability when used under defined conditions in the tropics and subtropics. Alley cropping involves the planting of crops in alleys or interspaces between repeatedly hedgerows of fast-growing, preferably leguminous, woody perennials. With the inclusion of more climatic and socioeconomic data and improved advisory recommendations, UNU-AES can be expanded to provide advice on alley cropping in more diverse geographical and ecological conditions and eventually address other agroforestry techniques. In 1991, serious efforts have been started in Egypt to develop crop management expert systems for different crops. A prototype for an expert system for cucumber seedlings productions has been developed. This prototype has six functions: seed cultivation, media preparation, control of environmental growth factors, diagnosis, treatment, and protection. The implementation has used the Hypertext facility included in the EXSYS shell. The overall control was implemented using the language provided by the tool, and consequently, the rule base was divided into modules according to the system functions (Rafea et al., 1991). Another expert system for cucumber production management under plastic tunnel (CUPTEX) was developed to be used by the agricultural extension service within the Egyptian Ministry of Agriculture and by the private sector. The main objective of developing such systems is to transfer new technology in agromanagement to farmers through packaging this technology using expert systems. This will lead to increasing the production and hence the national income, on the one hand, to reducing the production cost, on the other hand. CUPTEX is composed of three expert systems for fertigation, plant care, and disorder remediation. The validation of CUPTEX has revealed that it over-performs a group of experts in different specialties. Although the direct payoff is not the main objective of the development, experimentation with CUPTEX in two research sites has revealed an approximately 26% increase in the production and a 15% decrease in direct cost. These results show that CUPTEX could pay off approximately 60% of the development cost in 4 months if it is used to manage the plastic tunnels cultivating cucumber in the five sites in which it was deployed (Rafea et al., 1995). This system is the first deployed agricultural expert system to be selected in the conference of innovative application for artificial intelligence. It is also the first deployed expert system in developing countries -- not only in agriculture, but in other fields as well. In Italy, an expert system for integrated pest management of apple orchards (Gerevini et al., 1992), POMI, has been developed. Integrated pest management is definable as the reasoned application of agronomic methods products as well as chemicals, to allow optimal productive factors, while respecting the farm worker, the environment, and the consumer. POMI addresses the first preliminary phase of the complex process of apple orchard integrated control, namely the detection of the insect populations in the field and approximate the populations dimensions. The system consists of two parts: classification of user findings and explanation of these findings using abductive reasoning. In India, an intelligent front-end for selecting evapotranspiration estimation methods (Mohan and Arumugam, 1995) has been integrated with methods that calculate the evapotranspiration factor. Evapotranspiration (ET) is an important parameter needed by water managers for the design, operation and management of irrigation systems. Since there are many methods to compute ET, based on climatic data, an inexperienced engineer or hydrologist is perplexed with the selection of an appropriate method. An intelligent front end expert system (ETES) has been developed to select suitable ET estimation methods under South Indian climatic conditions. Ten meteorological stations located in different climatic regions and thirteen ET estimation methods have been considered in this ES. Along with the recommended method, ETES also suggests suitable correction factors for converting the resulting ET values to those of methods that result in accurate estimation. A picture-based expert system for weed identification (Schulthess et al., 1996) has been developed. Most current expert systems for weed identification are rule based and use text that contains a large number of botanical terms. In this system, the hierarchical classification generic task was used and the text descriptions were replaced with pictures to minimize the use of technical terms. Hypotheses are established or ruled out on the basis of the user's choices among options presented as pictures.
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |