=Paper=
{{Paper
|id=None
|storemode=property
|title=Urban land use planning using answer set
programming - preliminary report -
|pdfUrl=https://ceur-ws.org/Vol-1659/paper11.pdf
|volume=Vol-1659
|authors=Jennifer Muñoz,Flavio Everardo
|dblpUrl=https://dblp.org/rec/conf/lanmr/MunozE16
}}
==Urban land use planning using answer set
programming - preliminary report -==
https://ceur-ws.org/Vol-1659/paper11.pdf
Urban Land Use Planning using Answer Set
Programming
- Preliminary Report -
Jennifer Muñoz1 and Flavio Everardo2
1
Universidad Popular Autónoma del Estado de Puebla
and Instituto de Ciencias de la Benemérita Universidad Autónoma de Puebla
jenniferjocelin@hotmail.com
2
Universidad de las Américas Puebla
flavioeverardo@gmail.com
Puebla, Mexico
Abstract. The urban planning is required to direct the cities develop-
ment based on the sustainable use of its territory. On the other hand,
Answer Set Programming (ASP) is a logic programming paradigm with
the ability of non-monotonic reasoning, which has contributed to solving
complex problems in a short period of time, providing a considerable
number of possible outcomes to the same problem. The treatment of
sustainable urban planning by using a declarative approach like ASP
has not been proposed so far. This article aims to demonstrate a mod-
elling exercise using basic knowledge about a mix of urban land usage.
Also, to show alternatives for a better functional integration of green and
habitational areas, as well as a discussion of further research using this
approach.
Keywords: Urban Planning, Urban Land Usage, Answer Set Program-
ming, Knowledge Representation and Reasoning
1 Introduction
The need to think about an ordered cities growth and their development, are
not new topics. The proposal on strategic measurements to improve the condi-
tions of cities, is one of the issues considered in the Agenda 21, to promote the
construction of an international sustainable model for the twenty-first century
[13].
Derived from the above, with support from the United Nations Environment
Programme (UNEP), different countries worldwide have joined forces in order
to adapt environmental policies into their development processes with a sustain-
able approach. The Institute of Hygiene, Epidemiology and Microbiology of the
Ministry of Public Health of Cuba in coordination with the World Health Or-
ganization (WHO) and UNEP through the Latin American and the Caribbean
Regional Office, have proposed a classification of six universal principles for plan-
ning a sustainable urban ecosystem [19]:
82
1. Ensure proper water supply.
2. Keep protected vegetation.
3. Preserve soil quality.
4. Ensuring sustainable conditions for wildlife.
5. Keeping potential of regional food production.
6. Create a human scale urban environment.
Taking into consideration the principles 2,3 and 6, this article proposes form-
ing a basic knowledge that serves an intelligent system in resolving a complex
combinatorial problem such as urban land use planning. Initially this article
mentions works that use Knowledge Representation (KR) as a support in the
field of organization of urban territories, then the land use term is defined as
part of sustainable urban planning and an introduction to ASP is provided. Con-
secutively both issues are linked to generate a proposal where ASP contributes
to the definition of urban land use. This paper concludes with a discussion of
the preliminary results and future work.
2 Backgrounds
Land use and urban planning in general have been areas of interest to the scien-
tific community. Reason why, they have developed works that employ technology
as a way to propose alternatives for cities improvement. For purposes of this pa-
per, the works mentioned below have used some method of KR.
One of the first works is the URBYS system, defined as an expert system for
urban areas analysis through KR for decision-making [16]. There are studies that
address the selection of industrial zones [17] under a system called MATISSE
which, by decision tables represents knowledge as a set of rules and their respec-
tive actions. Likewise decision tables have also been used to introduce functional
classification theory to land use planning [18].
Changing the subject, ontologies is another used and proven approach in
works like OSMoSys [12] which seeks to harness the potential of Web applications
as interactive tool gathering information from urban areas as well as information
generated by humans through smart devices and social platforms.
It should be noted that these examples have used different approaches of AI
to resolve urban problematics. Nevertheless, so far none has taken advantage
of the benefits of ASP to solve these problems and that is why the strategy
proposed in this article is unique in its class.
3 Land Use within Sustainable Urban Planning
A standard definition of land use is the employment that humans make the
earth’s surface, covering the management and modification of the natural envi-
ronment to make it a built environment3 . Examples are human settlements that
3
https://en.wikipedia.org/wiki/Land use
83
involve industrial parks, shopping centers, roads, green and residential areas,
among others.
Within the universal principles for sustainable urban ecosystem planning
previously mentioned, three of them are directly linked to land use planning
purposes considering the functional integration between green and residential
use areas [8].
This work considers the functional integration of green and residential areas
as part of land use planning in a city. Reasons why the principles, protected
vegetation, soil quality and human scale urban environment are interrelated.
The fact that exists areas with trees, keep soil and air quality and this kind of
conditions are attractive for living. These three principles handle certain rules,
restrictions, conventions or preferences that allow forming a knowledge base for
modeling desirable behavior in order to adapt to conventional urban spaces.
4 Answer Set Programming
Several researchers have proposed the use of ASP as a solution to the paradigm
of Knowledge Representation and Reasoning (KRR) [2] [7] [9] [10], also since
its inception, ASP has been considered as the most suitable and effective com-
putational processing mode for non-monotonic reasoning and KR [2] [3] [11].
This perspective has promoted the creation of high performance and efficiency
Answer Set Solvers.
ASP has been tested successfully in areas, which involve problems of combi-
natorial search using a considerable amount of information to process [7]. It is
noteworthy that at the University of Potsdam has developed software as Clasp,
Gringo [5], Clingo [6], among others4 , which have been recognized worldwide for
its capacity and performance applied in ASP [1] [4].
Additionally, ASP is based on a simple yet expressive rule language that
allows users to easily model problems in a compact form. The solutions to such
problem are known as answer sets or stable models [15].
4.1 Problem Representation in Logic Form
To properly represent a problem it is necessary to segment it into three logical
components: Facts and Constraints, Preferences and Real Scenarios or Input:
• Facts and Constraints: are those foundations that we can get from a specific
problem whether they are not going to change or are prohibited. In other
words are the guidelines and parameters involving sustainable urban plan-
ning.
• Preferences: Establish rules and restrictions about situations preferable over
others [14].
4
More information about the Potsdam Answer Set Solving Collection software:
http://potassco.sourceforge.net
84
Modeling rules, constraints and preferences give us the parameters of what is
desirable, undesirable and preferably having and not having to solve a problem
based on the input. This information will be extracted from literature and real
sustainable cases (RSC) to become KB (Fig1).
Fig. 1. Information extraction to create KB
• The Real Scenario or Inputs: refers to the current situation once modeled
the problem. This needs to include a well-detailed description or specific
characteristics of a determined urban center. This set will call R (Fig2).
Fig. 2. Real or fiction case extraction to create R
If R changes, the result will be different even if KB does not change. The
benefits of ASP allow us depending on the complexity of the problem, is possible
to have a wide range of perspectives for decision-making from the same R.
85
5 Land Use Representation in ASP
Once represented a complex problem as specified in the preceding paragraphs, it
is necessary to translate this information into logical language for ASP. Taking
as example urban land usage, there may be more than 3 or 4 ways to grow a
city or use lands in a proper way.
Fig. 3. Example of houses and green areas distribution by KB definition
The Figure above (Fig3) shows an example of an area which it has been
demonstrated to be appropriate to build residential houses and green areas, but
the field should contain one-third of green areas at all times. In other words if
we have an area that can store 6 buildings, 2 of these must be green areas, the
other 4 must be houses.
To represent KB we can define two rules such as: 1. In each position is possible
to build whether a house or a green area. 2. Desirable number of green areas is
equal to the total space divided by 3.
Listing 1.1. Basic Land Use Example in ASP
1 1{ h o u s e I n P o s i t i o n (P) , g r e e n A r e a I n P o s i t i o n (P)}1: − p o s i t i o n (P ) .
2 1 { houseType (C, P) : house (C) } 1 :− h o u s e I n P o s i t i o n (P ) .
3
4 housesNumber (N) :− N = #count { h o u s e I n P o s i t i o n (C) } .
5 greenAreasNumber (N) :− N = #count { g r e e n A r e a I n P o s i t i o n (C) } .
6
7 desiredNumberOfGreenAreas (A/ 3 ) :− a r e a (A ) .
8 :− greenAreasNumber (N) , desiredNumberOfGreenAreas (NAV) ,N!=NAV.
9
10 b u i l t ( house , TC, P) :− houseType (TC, P ) .
11 b u i l t ( greenArea , P) :− g r e e n A r e a I n P o s i t i o n (P ) .
The Basic Land Use Example code above shows the set of rules and restric-
tions that arises for KB where lines 1 and 2 talk about the rule 1. Lines 4 and
5 carry the counting houses and green areas. Line 7 calculates the total of green
areas to have by dividing the total space by 3 and line 8 restricts that the num-
ber of green areas were different than the calculated. Finally, lines 10 and 11
indicate where the houses and green areas will be built. On the other hand it is
86
necessary to define the current scenario or R1 defining a first example indicating
that the area is of size 6 and there is only one kind of house to be built.
Listing 1.2. Basic Land Use Example for R1
1 area ( 6 ) . %% Area with 6 s l o t s f o r c o n s t r u c t i o n
2 p o s i t i o n ( 1 . . A) :− a r e a (A ) . %% T h e r e f o r e t h e r e a r e 6 p o s i t i o n s
3 house ( 1 . . 1 ) . %% E x i s t 1 type o f house
4
5 %% C o n s t r a i n t t h a t i n t h e 5 th p o s i t i o n a house must be b u i l t
6 %% This can be a bad q u a l i t y l a n d
7 :− b u i l t ( greenArea , 5 ) .
If only using the first three lines, we can see that there is a total of 15 solutions
combining in the 6 spaces, the number of houses and green areas respecting
the corresponding number of green areas. Adding the constraint that the area
number 5 is not suitable for a green area construction, so a house must be built.
This may satisfy certain qualities of the land mentioned in the previous section.
Running this example will deliver a total of 10 solutions. This amount of results
could simplify decision-making, but it is possible that in another scenario R2,
there are three different models of houses to be built. The complexity of this issue
has been raised for a total of 1,215 viable results. It is worth to mention that if
we add the same constraint listed above, the number of results will decrease to
810.
Listing 1.3. Basic Land Use Example for R2
1 area ( 6 ) . %% Area with 6 s l o t s f o r c o n s t r u c t i o n
2 p o s i t i o n ( 1 . . A) :− a r e a (A ) . %% T h e r e f o r e t h e r e a r e 6 p o s i t i o n s
3 house ( 1 . . 3 ) . %% E x i s t s 3 t y p e s o f h o u s e s
Finally, if later we found a real scenario like R3 expressing a land where fits 9
buildings instead of 6, and applying the same KB, the total results exponentially
increase to 61,236. Again, by adding the mentioned constraint we can limit the
result to 40,824.
Listing 1.4. Basic Land Use Example for R2
1 area ( 9 ) . %% Area with 9 s l o t s f o r c o n s t r u c t i o n
2 p o s i t i o n ( 1 . . A) :− a r e a (A ) . %% T h e r e f o r e t h e r e a r e 9 p o s i t i o n s
3 house ( 1 . . 3 ) . %% E x i s t s 3 t y p e s o f h o u s e s
In summary, KB has not changed over time but R has exemplified three
different cases and the results show an exponential growth leading to an extensive
analysis before a decision can be made. The more complex the problem, the more
results we can get. It is worth noting that the greater the details modeled in R
including constraints, more specific results we can get, limiting an exponential
growth and making easier the analysis of the results and decision-making.
87
6 Discussion
In this paper is presented a new and innovative approach to begin to contribute
to the understanding of the urban structure and planning with the help of ASP.
Also it is demonstrated that a declarative programming perspective fits properly
in the urban problems modeling.
It is worth mentioning that the set of characteristics of an urban society is
manifested in its territorial structure, therefore, each city will have a different soil
use structure [8]. This article does not include a detailed description of the spatial
structure of an urban center in particular; reason why it is planned for future
research, enrich KB so it is enough to adapt to different escenarios. In order to
find the best plan for land use structuring or restructuring, considering the needs
of functional integration for a specific urban ecosystem, and also, respecting the
methodology and problem-solving strategy presented in this article. Hence is
considered possible to explore and venture later in the 6 universal principles for
sustainable urban planning.
Finally, once advanced the following works where both KB and R are in a
higher level of maturity, we will seek to test four hypotheses, which are:
• Review a real scenario R to determine whether or not this particular zone
makes a proper use of lands.
• Recommendations or modifications for R with the purpose of finding alterna-
tives to fulfill a proper land use management.
• Projections of how R can be developed in short, medium and long term.
• Forecasts that show how R could grow if certain changes or adjustments on
prior knowledge are not made.
References
1. Mario Alviano, Francesco Calimeri, Günther Charwat, Minh Dao-Tran, Carmine
Dodaro, Giovambattista Ianni, Thomas Krennwallner, Martin Kronegger, Jo-
hannes Oetsch, Andreas Pfandler, et al. The fourth answer set programming com-
petition: Preliminary report. In Logic Programming and Nonmonotonic Reasoning,
pages 42–53. Springer, 2013.
2. Christian Anger, Kathrin Konczak, Thomas Linke, and Torsten Schaub. A glimpse
of answer set programming. KI, 19(1):12, 2005.
3. Chitta Baral. Knowledge representation, reasoning and declarative problem solving.
Cambridge university press, 2003.
4. Francesco Calimeri, Giovambattista Ianni, and Francesco Ricca. The third open
answer set programming competition. Theory and Practice of Logic Programming,
14(01):117–135, 2014.
5. M. Gebser, T. Schaub, and S. Thiele. Gringo: A new grounder for answer set
programming. Logic Programming and Nonmonotonic Reasoning, pages 266–271,
2007.
6. Martin Gebser, Roland Kaminski, Benjamin Kaufmann, and Torsten Schaub.
Clingo= asp+ control: Preliminary report. arXiv preprint arXiv:1405.3694, 2014.
88
7. Michael Gelfond and Yulia Kahl. Knowledge representation, reasoning, and the
design of intelligent agents: The answer-set programming approach. Cambridge
University Press, 2014.
8. Ignacio Kunz. Usos de suelo y territorio: tipos y lógicas de localización en la ciudad
de México. Plaza y Valdés, 2003.
9. Nicola Leone, Gerald Pfeifer, Wolfgang Faber, Thomas Eiter, Georg Gottlob,
Christoph Koch, Cristinel Mateis, Simona Perri, and Francesco Scarcello. The
dlv system for knowledge representation and reasoning. 2002.
10. V. Lifschitz. What is answer set programming. In Proc. of AAAI, pages 1594–1597,
2008.
11. Ilkka Niemelä. Logic programs with stable model semantics as a constraint pro-
gramming paradigm. Annals of Mathematics and Artificial Intelligence, 25(3-
4):241–273, 1999.
12. Achilleas Psyllidis. Ontology-based data integration from heterogeneous urban
systems: A knowledge representation framework for smart cities. In CUPUM 2014:
Proceedings of the 14th International Conference on Computers in Urban Planning
and Urban Management, Cambrigde, USA, 7-10 July 2015. MIT, 2015.
13. Nicholas A Robinson, LE Astorga, T Oksanen, EJ Trigo, O Pastora, P Pastor, S de
Izaguirre, AC Padgett, R Arita Arita, H Palma, et al. Agenda 21: earth’s action
plan. Number GTZ-798. IUCN, Gland (Suiza)., 1993.
14. Torsten Schaub. Preferences in answer set programming.
15. Patrik Simons, Ilkka Niemelä, and Timo Soininen. Extending and implementing
the stable model semantics. Artificial Intelligence, 138(1):181–234, 2002.
16. Emile Tanic. Urban planning and artificial intelligence: the urbys system. Com-
puters, Environment and Urban Systems, 10(3):135–146, 1986.
17. Frank Witlox. Matisse: a relational expert system for industrial site selection.
Expert systems with applications, 24(1):133–144, 2003.
18. Frank Witlox, Marc Antrop, Peter Bogaert, Philippe De Maeyer, Ben Derudder,
Tijs Neutens, Veronique Van Acker, and Nico Van de Weghe. Introducing func-
tional classification theory to land use planning by means of decision tables. De-
cision Support Systems, 46(4):875–881, 2009.
19. Annalee Yassi, Tord Kjellstrom, Theo deKok, and Teco Guidotti. Salud ambiental
básica. Oficina Regional de Cultura para América Latina y el Caribe, 2002.
89