The development of expert systems in agriculture consists of many steps such as problem definition, selection of experts, audience considerations, knowledge representation, coding, testing, and feedback. The problem definition and selection of experts for the problem domain are the foundation of a working system. Audience definition, economics and goal setting are areas that must be documented before knowledge engineering. Knowledge representation methods and system conceptual layout are the next level of development. The use of the user feedback and field testing data to improve the system are often overlooked. Benefits of expert systems for on farm decision making include education, efficiency, and adaption to changing regulations. Many aspects of agricultural expert systems are similar to traditional expert systems; yet special problem inherent in agriculture make the development interesting and challenging.
Janice E. McClure
A.P. Papadopoulos, J.L Shipp, W.R. Jarvis, T.J. Jewett, and N.D. Clarke
Greenhouse crop production technology is advancing rapidly, and the management of greenhouse crops has become increasingly difficult. Computerized environment and fertigation control of greenhouse crops grown in soilless media offer opportunities for unparalleled manipulation of crop growing conditions. However, the optimization of crop growing conditions for maximum productivity must be practiced with an eye on environmental regulations; worker health concerns; consumer demands for safe food; and ultimately on energy, water, fertilizer, and pesticide use economy. Managing the complex greenhouse cropping system requires a multidisciplinary approach that integrates pest and disease protection strategies with routine cultural practices and environmental and fertigation regimes into a common decision-making process or Integrated Crop Management strategy. This poster describes an Expert System for greenhouse cucumber management based on a general model of Integrated Crop Management for greenhouse crops.
Glenn H. Sullivan, William J. Ooms, Gerald E. Wilcox, and Douglas C. Sanders
1 Professor, Marketing Economics. 2 Systems Engineering Economist. 3 Professor, Plant Nutrition. 4 Professor, Horticulture. Purdue Univ. Agricultural Experiment Station Journal no. 12645 The cost of publishing this paper was defrayed in part
K. Bergsma, S. Sargent, J. Brecht, and R. Peart
Temperature management is the most widely used method to extend the postharvest life of vegetables. Unfortunately, during less than optimal commercial conditions, certain commodities can be exposed to low, nonfreezing temperatures that may shorten their market life due to chilling injury (CI). CI is difficult to diagnose since not all commodities exhibit the same symptoms. Environmental factors may also affect the expression of CI The services of an expert are usually required to positively diagnose CI, however, experts are not always readily available, particularly during routine commercial handling. An expert system, a computer program that emulates a human expert's thought processes, will be developed to diagnose CI symptoms for several commodities. A prototype developed with Level5 Object, an expert system shell, will be presented. Diagnosis is determined by applying rules and certainty factors based on user responses to queries on the type and extent of visual symptoms. The applicability and advantages of this system will be discussed.
R. M. Crassweller, J. W. Travis, P. H. Heinemann, and E. G. Rajotte
Apple orchards are highly diversified and complex ecological and economic systems. Production is affected by a wide range of insects, diseases, weeds, and mammalian pests. The incidence of these pests is often dependant upon climatological effects; and the microclimate within orchards. An expert system, a form of artificial intelligence, has been developed and commercially released to apple growers that utilizes weather data to make recommendations regarding production decisions. Users of the system are instructed on how to establish a weather station, and to collect, and input weather data from the farm. The information is utilized to calculate disease infection periods and pesticide residues to arrive at a control recommendation. Other weather dependant modules include the scheduling of trickle irrigation as well as water application rates during a frost. An interactive demonstration of the system will be presented to the group.
Edward F. Gilman
This computer program, delivered-on a CD-ROM disc, develops a list of tree species and cultivars suited for a specific planting site. It requires little previous computer experience or tree knowledge to operate. Using multiple choice questions, the program automatically brings the user through above ground and below ground site analysis. This includes all the considerations known to influence proper species section for a planting site. Using C++ programming and the NASA-developed expert system shell called CUPS, a list of facts is generated as the user answers the questions. At the press of a button, the program finds trees that match the attributes the expert system placed on the facts list. The list can be further modified by choosing among ornamental and other tree attributes that might be of interest to the user. The tree list can be printed in several seconds. A typical run through the expert system takes 2 to 4 minutes to answer about 20 to 25 questions. The program contains data on 681 trees, more than 1,800 color photographs, and a 4-page fact sheet including 3 line drawings for each tree totaling more than 2,000 pages. The program can also be used as a reference by paging through the tree records to find information about specific trees. Each tree record lists on the computer monitor a large variety of data for the tree, allows you to view text about the tree, displays a line drawing of the entire tree, and displays up to seven photographs of each tree. The program will be distributed nationwide as a tool to help landscape architects, horticulturists and others select the right tree for the right place.
R.M. Crassweller, J.W. Travis, P.H. Heinemann, and E.G. Rajotte
Decreasing resources and increasing complexity of horticultural crop production necessitate that new technologies be developed to transfer information to commercial producers. Expert systems (ES) have been cited as potential tools that can facilitate knowledge transfer. The definitions of an expert system, however, technically only indicates a computer program that simulates the thought processes of a human expert and, as such, does not supply all the facets necessary to assist commercial producers. The combination of databases, graphic capabilities, and textual information into a comprehensive program would provide a more complete package. To differentiate the two, we use the term decision support systems (DSS). The development, testing, and release of DSS, however, require greater commitment and interdisciplinary cooperation. Developing DSS fosters interstate, interregional, and international cooperation among researchers and extension personnel. Using systems developed in fruit production as examples, we outline the value of DSS to promote cooperation, the resources necessary to develop these systems; and the attitudinal change necessary to build the systems.
R. E. Plant, F. G. Zalom, J. A. Young, and R. E. Rice
An expert decision support system for agricultural management called CALEX is currently being developed. The program runs on any IBM compatible personal computer with 256K or more of memory and either two floppy disk drives or a hard disk and one floppy disk drive.
Robert M. Crassweller, Paul H. Heinemann, and Edwin G. Rajotte
Apple orchards are highly diversified and complex ecological and economic systems. Production is affected by a wide range of insect, disease, weed, and mammalian pests, and is subject to the same economic and social constraints as any business enterprise. A computer technology, expert systems (ES), is being used to assist fruit growers, county extension agents, and private consultants in making better decisions about the complex horticultural, entomological, and pathological orchard problems.