Fuzzy expert system for land reallocation in land consolidation

Abstract

One of the most important steps of land consolidation projects is land reallocation studies. In Turkey, reallocation studies carried out in the scope of land consolidation projects are made according to farmer preferences (interviews). In addition to interview-based land reallocation model, mathematical models have been used in the previous optimization studies for reallocation procedure. Recently, fuzzy logic method, which is capable of modeling human mindset and used when other forms of mathematical models cannot be developed, has also been applied to the field of geomatic engineering, as well as in other engineering branches.

This study examined the applicability of a fuzzy logic method at the reallocation stage of land consolidation study, where development of an accurate mathematical model was not possible. The results obtained from the fuzzy logic-based land reallocation model were compared with those obtained from the interview-based land reallocation model. Farmers were surveyed to determine which land reallocation model they preferred. The results indicate that 80.5% of the participant landholdings were satisfied with the fuzzy logic-based reallocation land model, while 50% were with the interview-based land reallocation model.

Introduction

Land consolidation (LC) is a tool for improving the effectiveness of land cultivation and for supporting rural development (Sklenicka, 2006). As an important approach to achieve the sustainable utilization of land resource, land consolidation not only need to regard the amount of the farm land for the sake of achieve thing homeostasis of farmland, but also need display the active effect in other aspects, such as improve the quality of farmland, reform the ecological condition and promote the adjustment of the economic formation (Zou et al., 2008).

In the whole countries, land consolidation is applied to improve the rural areas. Because rural areas comprise substantial parts of the regions and are subject to a range of pressures including water shortage, land degradation, failing commodity prices and depopulation. Land consolidation means to unite and reregister the lands, which were divided because of heritage, sales or irrigation canals (Cay & Iscan, 2004).

In Western Europe, for example, in Germany and in the Netherlands, LC is often a part of a wider regional development programme for rural areas. In those regional development programmes LC is used for enhancing systematic land use in the rural areas and for readjusting the areas according to the assignment of the programme (Vitikainen, 2004). The contents of the LC process include similar main stages in all Europe. The process consists of the preparation, the inventory, planning, and the implementation stages, each varying in extent and duration.

In Turkey, 8.5 million ha (out of an arable area of 28.5 million ha) can be irrigated economically, but, at present, only 4.8 million ha are being irrigated. The average farm size was 10 ha in 1950, 6.8 ha in 1980, 5.9 ha in 1990, and 6.1 ha in 2001; the numbers of farms in the same years were 2.2 million, 3.5 million, 3.9 million, and 3.02 million, respectively (Anon, 2004, Babagiray, 2006, Gun, 2003). The average parcel number per landholding, although different from region to region, is 4.08, according to the results of the General Agricultural Census in 2001. This is equivalent to 1.50 ha/parcel (i.e. 6.1 ha/4.08 parcel) in 2001.

Land reallocation is the most important and a time-consuming stage of land consolidation studies since high number of criteria are considered at this stage. It is crucial for social peace to conduct land reallocation studies in such way to meet the demands of farmers and also the principles of equity and justice.

The problem encountered in land consolidation studies can be defined as allocating “n” number of cadastral parcels to “m” number of blocks. To this end, optimization studies have been conducted for land reallocation process, which are based on many mathematical models (Avci, 1999, Ayranci, 2009, Girgin and Kik, 1989, Kik and Sprik, 1990, Lemmen and Sonnenberg, 1986). However, many different solutions have been suggested as there is no single accurate mathematical model for the land reallocation process. However, since linguistic statements and human considerations that affect the reallocation could not be embedded in them, these mathematical models have been reported to achieve low success.

Since it is capable of incorporating human experiences that can be expressed linguistically but are difficult to express mathematically, fuzzy logic (FL) method can be utilized at the reallocation stage of land consolidation projects. In engineering and other disciplines, events and systems are defined by using accurate mathematical models. These models are taken as basis in the estimations about future status or behavior of the event or system. However, majority of daily problems or situations cannot be either fully modeled due to various reasons or may not refer to a specific event. The FL approach can be utilized in analyzing and solving such problems.

The origin of the FL approach dates back to 1965 since Lotfi Zadeh’s introduction of the fuzzy set theory and its applications. Since then the FL concept has found a very wide range of applications in various domains like estimation, prediction, control, approximate reasoning, pattern recognition, medical computing, robotics, optimization and industrial engineering, etc. (Sen, 2004).

Zadeh (1965) published his famous paper “Fuzzy sets” in Information and Control providing a new mathematical tool, which enables us to describe and handle vague or ambiguous notions such as “a set of all real numbers, which are much greater than 1”, “a set of beautiful women”, or “a set of tall men”. Since then, fuzzy set theory has been rapidly developed by Zadeh himself and numerous researches, and an increasing number of successful real applications of this theory in a wide variety of unexpected fields have been appearing in open literature. The main idea of fuzzy set theory is quite intuitive and natural. Instead of determining the exact boundaries as in an ordinary set, a fuzzy set allows no sharply defined boundaries because of generalization of a characteristic function to a membership function (Sakawa, 1993).

The framework of fuzzy logic is unique in its ability to represent subjective or linguistic knowledge in terms of a mathematical model. For this reason, FL provides a natural method for constructing systems that emulate human decision making processes. Literature on the subject of FL systems is extensive and applications, particularly in the field of fuzzy control and fuzzy expert systems, are prevalent. Mendel, 1995, Klir and Yuan, 1995 provide good introductory texts on FLSs, while some examples of applications of FLSs may be found in Sugeno and Park, 1993, Maiers and Sherif, 1985, Kandel, 1991, Ramot et al., 2003.

FL is a recognized instrument for modeling in many scientific and technical fields. There are also a lot of problems where fuzzy methods can be used to reach better solutions than classical models can do. It concerns on the one hand questions, where uncertain parameters occur, which cannot be handled by classical methods in adequate way. On the other hand, there are problems where linguistic fuzzy rules can describe relations better than it can be done by crisp mathematical formulas.

A number of previous studies have been carried out in the field of geomatic engineering and on FL and artificial neural networks. Heine (1999) successfully used artificial neural networks (ANN) and FL systems while modeling body deformation of a dam. Miima and Niemeier (2004) used feed-forward ANN to accurately model the movements of a historical stone bridge over a river. Akyılmaz, Çelik, Apaydın, and Ayan (2004) modeled vertical movements of Fatih Sultan Mehmet Bridge, using atmospheric and bridge traffic density data, as well as GPS observations and feed-forward ANN. Pashaev, Sadykhov, Yildiz, and Karabork (2005) used artificial neural networks (ANN) in simulate surfaces and map geofield parameters. Akyılmaz (2005) produced numerical solutions to the problems of forecasting of the Earth’s rotation parameters, determination of earth’s gravity field through artificial satellite data and computation of Geoid through GPS-Leveling; analyzed these solutions and; made comparisons between these analysis from different aspects. Using ANN, Barsi (2001) achieved rotation at sufficient accuracy level, between coordinates defined at different data and different projections in smaller fields. Akyılmaz, Erden, and Ipbuker (2005) introduced a transformation model for the calculation of ellipsoidal latitude and longitude on the basis of projection coordinates, by using fuzzy inference systems. They used Ginzburg IV projection, also known as the CNIIGAIK 1939–1949 projection, which was a former Soviet projection designed in tables and used for mapping the whole world. Key et al., 1989, Wang, 1990, Kanellopoulos et al., 1992, German and Gahegan, 1996, Thackrah et al., 1999 published studies on remote sensing based on FL. Recently, FL has also been applied to image processing (Bezdek, 1981, Karmakar Gour and Dooley Laurence, 2002, Kundu and Chaudhuri, 1993, Looney, 2000, Nakamuro, 1996, Pal et al., 1983, Pal and Rosenfeld, 1988, Russo and Ramponi, 1995, Tizhoosh and Michaelis, 1998). In his doctoral thesis, Ustuntas (2005) suggested a fuzzy logic-based image matching algorithm.

In the present study, FL was applied at the reallocation stage of a land consolidation study, which could not be supported with an accurate mathematical model. Ilgın-Agalar part of Konya Province (Turkey) was selected as e study site. Local residents were interviewed to learn their ideas about land allocation. The results of the interview-based land reallocation model and fuzzy logic-based land reallocation models were compared. Comparison criteria were chosen as follows: The number of parcels and shares; average size of parcels; average number of parcels per landholding; development times of new parceling plans; project cost; the relationships of landholdings with their close relatives (partner, father, mother, siblings and other landholdings whose land s/he uses), and; to what extent do the results comply with the demands listed by the farmers during the interviews. In addition, a questionnaire was prepared to establish farmers’ preferred land reallocation model.