A Knowledge Model And Decision Support System For Processing Tomato Management

Processing tomato(Lycopersicon esculentum Mill)in Xinjiang arid area is superiority characteristic industry, and holds the high status in the agricultural development strategy. After the development of the past few years, processing tomato has become the second largest economy crop. At present, the processing tomato planting in arid zone has been initially entered the industrial development of the road, but how to achieve balanced production and the modernized cultivation and management, is the critical technical issues to which the tomato production units would face. To solve this problem, the development of decision support system for processing tomato management is the key step.In this study, based on the knowledge model construction principle, the author applied the system analysis principal and mathematic modeling technique to study the knowledge expression system for processing tomato cultivation management, based on the collection and use of the knowledge and data of the processing tomato planting theory and technical and experts' experience, with the support of necessary experiment, after analyzing, generalizing and synthesizing the relationship among processing tomato growth character, management indexes, variety types, ecological environment and production level, by further system analysis and quantitative expression by mathematical and statistical methods, a knowledge model for processing tomato management with temporal and spatial characters, which based on the relation of crops, circumstance and measures, was established. By further expression system for processing tomato management, a comprehensive integrating the knowledge base and intelligent knowledge model decision support system for processing tomato management (KMDSSPTM) was established.In the dynamic knowledge model system, there were two parts as cultural technique plan design and suitable dynamic growth index prediction. The knowledge model for cultural technique plan design includes submodels of target yield calculation, variety selection, sowing or transplanting date, population density and seeding rate, fertilization strategy, water management. The knowledge model for the dynamics of main development indices includes submodels of suitable development stages, plant height, leaf area index, dry matter accumulation, and nutrient accumulation status for aboveground plant.The submodel for target yield prediction was dynamic quantized based on the gradually revised model of"crop growth dynamic statistical methods". After calculating the highest radiation yield potential of decision-making point, the influence of temperature, water, soil fertility and cultivation management technology level to processing tomato yield was quantized, and based on the average yield of last three years, the estimated values of target yield were calculated; The submodel for variety selection was developed by quantifying the integrative effects processing tomato varieties characteristic value to environment factor and production demand with the principles of ecology and methods of multi-target decision. The appropriate and the latest sowing period were determined based on the arrangement of crop for rotation, variety maturity characters and development requirement to environment, by taking the mature period confidence as the weight standard and taking the quantitative physiological development time as the time scales to calculate the sowing date; The population density, according to the principle of determining branch number from target yield, population density from branch by introducing the variety type and branching parameters, and by integrating the effects of fertilizer and water management levels, through dynamic quantizing relation between different yield levels and suitable effective branching number, was obtained; Based on the planting density, after considering the effects of different soil types'physical-chemistry character and cultural practice on seedling emergence rate, the suitable sowing rate was calculated; Water management was based on quantization of relation between yield and water demand, to calculate total water requirement, then according to the principle of water balance, through considering water consumption rule, the total irrigation quota in entire development period of processing tomato and water allocation ratio in each development stage were reckoned up by obtaining the meteorological data including sunshine hour, temperature, humidity, rainfall amount to calculate field evapotranspiration and effective rainfall; The submodel for fertilization strategy, based on the principal of nutrient balance, was established by combining the effects of soil character, target yield and variety traits. In this submodel, the total fertilizer requirement was firstly calculated using the demanded quantity of nitrogen, phosphorus and potassium to meet the target yield requirement, and then according to the soil nutrient supplying and utilization ratio of fertilizer in that vary season to calculate the total requirement of nitrogen fertilizer, phosphorus fertilizer and potassium fertilizer. Finally according to the soil physical-chemistry character, yield level, nutrient consumption characteristics and fertilization method, the ratio of organic to inorganic nitrogen and the ratio of basal to top dressing fertilizer were determined.According to the principle of physiological effectiveness accumulated in the dynamic development being constant, through quantitative calculation of thermal and photoperiod effectiveness to physiological development of processing tomato, the intrinsic development factor was introduced and the dynamic prediction model of appropriate growth period was established; Using Monsi formulation and sunshine intensity in the decision-making point and photosynthetic performance of the basal leaves, combining the factors of target yield, variety traits and production measure, the largest appropriate leaf area index was firstly calculated, and on this bases the dynamic knowledge model for LAI of processing tomato, through relative physiological development time and normalized leaf area index, was then established; After established dry matter accumulation dynamic model by using increase rule of the Logistic curve quantified by harvest index and yield level, combining the sowing date and genotype of variety, the author developed the simulation model of dry matter partitioning and yield forecast which was based partitioning and harvest index; The submodel for nutrient index dynamic in processing tomato plant was driven by physiological development time and based on quantitative calculation of population nutrient accumulation and dry matter accumulation at main growth stages; The variation of plant height was described by quantizing the relationship between plant height in actual time and final plant height calculated by machine simulation of variety botanist's parameters.On the basis of the dynamic knowledge model for processing tomato management, a comprehensive and intelligent knowledge model-based decision support system for processing tomato management (KMDSSPTM) was developed with component technique and Delphi7.0 platform by incorporating the rule-based knowledge base system for processing tomato management. The KMDSSPTM realized the effective integration of the prediction and decision-making functions, provided a foundation for intelligent and digital management decision in processing tomato cultivation.Case studies on the knowledge model with data sets of daily weather data, variety types, soil types and so on in different eco-sites, such as Shihez, Yanqi, Alar, Zhangye and Bayanaor, indicated a good decision-making and wide applicability.