Comparative Analysis of Computational Models of Sustainable Transportation
Introduction
The problems associated with human development have become household conversations for all those in the developed world. Effective and sustainable transportation is at the root of many of these problems. Industrialized nations contribute 20-25 of CO2 emissions from motorized vehicles (TAO, HUNG, 2003). Motorized transportation, as it is now, is one of the leading causes of premature death (Yevdokimov 6). Countless other problems surround modern transportation methods and because of this the need for a newer, updated system of urban transportation has never been greater.
A preventative approach to designing a city should be adopted when traffic and mass transit are being designed. Enormous sums of money could be saved by adopting a precautionary attitude instead of a reactionary one. The public health of the population and efficiency by which people, goods and even information are transferred within a city would also benefit from a healthier, safer and ultimately more efficient transportation system.
Within the last 20 years computational models have been developed to address the issues of urban transportation systems. These tools define the indicators of sustainability as defined by the developer, set up acceptable ranges for the indicator, and organize the resulting information in a manner that is useful to the end user. The major difference in computational models comes from the development of the indicators and the methods by which they are organized but, as will be seen, defining what constitutes sustainability and the indicators that define good or bad sustainability is normally the most difficult part of creating and instituting a useful computational model. Often, the organization of the model changes the resulting methods the governing body uses to implement changes to their transportation system. The purpose of this paper is to analyze several indicative computational models that are being applied to improve urban transportation systems and compare them while also suggesting possible methods of improvement for each.
Before a computational model adapted for sustainable transportation can even begin to be drawn up the idea of sustainability must be defined as it is applied to transportation systems. The most commonly accepted definition of sustainability comes from the Brundtland Report, Our Common Future, Sustainable development is development that meets the needs of the present without compromising the ability of future generations to meet their own needs, (World Commission on Environment and Development, 1987). This definition is vague enough to be interpreted in a number of different ways but essentially comes down to meeting the three basic needs of any society, economic viability, social acceptance, and environmental safety. Several transportation computational models have adopted the
WCED definition of sustainability in developing an approach to traffic systems.
After the WCED several other ideas have been presented. The OECD conference in Vancouver in 1996 introduced a higher emphasis on the social aspect of sustainability in future models (Yevdokimov, 2003). Ultimately the Sustainable Transportation System (STS) by Yuri Yevdokimov was born out of this concept and is now one of the leading computational models utilized in Canada. The IUCN Renowned Thinkers Meeting in January 2006 and the International Symposium on Technology and Society 2009 have also formed their own definitions on how and what sustainability means to the modern world. From these, several modifications to already extant computational models were made such as the SUTRA and MARKAL models.
Where to Improve
The clich remark to make in respect to sustainability is to simply say that human activity should not affect the environment. Defining what that means, especially when it concerns people, is far more difficult. Various groups of scientists have approached the subject in different ways. A set of areas in which to improve the impact of humans must be laid out in any model. Some models focus on one area more than another but overall the list is generally the same. Land-use, greenhouse gas emission and possible energy alternatives along with current energy usage are some of the main environmental concerns that computational models address. In an increasingly interconnected environment, none of those areas can be looked at by themselves. Concerns of efficacy, transportation infrastructure, alternate modes of transportation, economics, and social acceptance must be considered in order to make a model useful.
Land-use for transportation takes an enormous amount of area. Zoning policies, commute distance and the number of vehicles on the road are huge contributors to the amount of space required to transport people. The total amount of land required is small when compared to the total amount of land available but they are concentrated in urban areas where their impact is more strongly felt (Yevdokimov, 2003). Along with the area used for actual roadways and parking, is the exact utilization of the land itself. For example, parks and recreational areas require less space per capita for roadways than do commercial areas. Suburban sprawl vs. urban compactness i.e. commute distance, needs to be assessed in a model as well (Fedra, 2004).
Figure SEQ Figure ARABIC 2. Description of Land used by transportation industry.
In order to improve transportation sustainability, the transportation infrastructure as a whole must be improved. Large amounts of money are being spent on infrastructure as people become more aware of the impact that vehicular traffic has on urban life. As the resources for transportation become smaller, developers are relying on ever more cost effective methods of solving transportation problems and are therefore relying more heavily on computational models to design their systems (Akinyemi, Zuudgeist 2002). To apply infrastructure to a model, it is defined as the various methods and systems of travel that are possible. For example mass transit either in the form of busses, trains and ferries vehicular traffic such as cars, trucks or even bikes and other modes of transport, cars, planes etc must be inserted in the model.
Traffic fatality and injury is a concern that is not as easily addressed by computational models but must be inserted nonetheless. Traffic fatalities measure in the thousands nationwide and should be considered a key aspect of every future transportation model. Many times, the design of a system of roadways can drastically reduce the amount of deaths and injuries an area reports.
Figure 3 Traffic fatalities per country as a measure of per capita and per distance traveled
Energy alternatives can be inserted into a model of transportation. Various methods of transport utilize various types of energy and have their own unique advantages and disadvantages. Mass transit in electric railcars tends to use less energy per capita. Electric cars can be considered, even the size of vehicles used by most citizens in a study on the sustainability of an area. Since decreasing the overall energy usage of the population is an obvious concern of the public it is included in most models.
Alongside energy usage as an area of improvement is that of emissions. Carbon dioxide, sulfur dioxide and particulate matter percentages within many city centers are far above the generally accepted safe levels. Hong Kong, for instance recorded an Air Pollution Index (API) of 153 on New Years Day of 2002. A level over 100 is considered very high by the EPA (2002). The number of cities with levels of unacceptable amounts of air pollution is increasing continuously. Up to 60 of all emissions can be attributed to vehicular traffic, despite many years of engineering efforts to reduce this amount (Akinyemi, Zuudgeist, 2002). It is now time that a model of effective transportation combat this problem and many others like it.
Indicators of a Computational Model
The indicators of a computational model are the exact identifiers by which a computational model will determine the sustainability of an area. They range from levels of carbon dioxide emissions to automobile travel patterns to population densities. As explained in the introduction, these indicators must be defined before a model can even be started. The body of people determining the indicators can have a major effect on the outcome of what is deemed important. They are the ones that determine what is and is not important to the subject of sustainability. The STS, implemented by the Canadian government but determined by several sets of bureaucratic bodies has a set of 14 sustainability indicators (Yevdokimov, 2003). The SUTRA project has a large number of individual indicators determined by the various models that fall under its umbrella, but groups them into five categories (Fedra, 2004). Similarly, the SILENT model groups its indicators into four categories (Yigitcanlar, Dur, 2010).
While keeping in mind that the three general objectives of any computational model are to appease the concerns of the social, economic and environmental aspects of the population a more specific set of sub-objectives must be determined. A list of six sub-objectives has been developed and is accepted by over 90 of policy makers in European cities (May et al, 2001, pp 12, 13)
economic efficiency
livable streets and neighborhoods
protection of the environment
equity and social inclusion
safety
contribution to economic growth
Further specification of this is list necessary, however before an effective model of sustainability can be built.
Two basic methods of defining the above sub-objectives exist, quantitative and qualitative. Qualitative models use sets of organizing principles to descriptively characterize transportation systems while quantitative models mathematically describe transportation systems (TAO, HUNG, 2003). For the purpose of this paper, quantitative computational models will be analyzed due to their use of objective, easily measurable indicators. A computational model of sustainability requires an extremely specific definition of each of the indicators that are deemed important by the designers. While the indicators for each model may vary they are essentially measured the same. Carbon dioxide levels are generally measured in the same methods. The differences lie in the importance, or weight, it is given and the organization the model uses in determining that importance. The same is true for nearly all of the indicators a computational model uses.
Examples of Computational Transportation Models
Three basic types of quantitative assessment models exist. Assessment indicator models, further divided into composite index models, multi-level index models and multi-dimension index matrix models are the most commonly used worldwide. System dynamics models utilize horizontal and vertical linkages to evaluate sustainability needs. Optimization models are the simplest of all the models, being comprised of basic single-line equations, often utilizing linear programming (TAO, HUNG, 2003)(Black et al, 2002). Optimization models, Systems dynamics models, and Assessment indicator models are widely used throughout the world market and because of that will be discussed in depth.
Optimization Models
A mathematical optimization model is a tool for addressing a single issue as it is constrained by social, economic and ecological factors (TAO, HUNG). The model is typically used to address a single issue at a time. Linear programming is the model of choice for many developers that choose to use this method. It is used extensively in many areas of problem-solving in engineering and finances and can be easily applied to sustainable transportation.
Just as with all models an indicator must first be defined. Travel distance, investment cost or heat island effect can be applied. Normally only one factor for each equation may be tested. An equation then derives a relative sustainability factor as it is applied to the three main objectives of sustainable transportation.
Linear programming can be used to minimize the amount of travel between two destinations (Black et al, 2002). This is the classic transportation problem. In this example, the original land-use destination is factored against the type of land traveled through, the distance traveled and finally the origin of the destination. An additional constraint excludes outlier trip flows. It is written as (Hitchcock 1941)
Figure SEQ Figure ARABIC 3
with xij travel from zone i to zone j, cij cost per unit of travel, Oi total travel possible per zone and Dj to total travel required for zone j (Hitchcock 1941). The end product gives a minimization value for cost of travel for various trip inputs. Trip distances as well as different destinations may be entered.
The result will be a cost analysis per unit of travel as compared to defined zones traveled. One result found is that commute distances from suburban neighborhoods tend to be longer than commutes within a centralized city center (Black 2002). Many other aspects of travel may be analyzed, but typically trip distance is set against an ecological andor economic constraint.
Optimization models as set forth by linear programming are an indispensable tool for analyzing a single aspect of transportation. They are severely limited, however, in the scope of their ability. All examples studied focus on a single issue within a broad scheme of criteria. For individual factors, usually trip distance, optimization models provide quick easily decipherable information. The difficulty comes from determining how those single issues are affected by or influence other aspects of sustainability. The above example focuses on two of the general areas of concern, economy and social land use. In fact, the only variable is the economy of trip distance as determined by the land use already in place. No input is available for the effect that changing land use policy will have on commute distance.
Environmental impact is not mentioned either. It is implied but not explicitly laid out. Trip distance may be the same for a person using mass transit, but the ecological impact will be less.
There are numerous problems with using linear programming to solve an issue as complex as sustainable transportation but it is still a useful tool. It presents information in a clearly defined mode. The information is then easily explained to those that really matter, the policy makers that will inevitably be changing the system. It is also possible to create a network of equations that can be manipulated to address all of the indicators as defined by the developers but this is typically not the case. Normally, governing bodies move towards a more complicated model that is more comprehensive.
Sustainable Transportation System, Canada
The Sustainable Transportation System (STS) was developed by Yuri Yevdokimov for the Canadian transportation system. TAO and HUNG (2003) define STS as a system dynamics model. The basic use of indicators to determine levels of sustainability in an area still exists, but the method by which the importance of each indicator is determined varies drastically from the assessment indicator models. The STS is much more complex than the optimization models because it addresses multiple criteria simultaneously.
Yuri (2003) reveals that the unique approach of the STS is the introduction of a normative aspect to the assessment of transportations sustainability. Instead of analyzing the three main objectives of sustainability as an island separate from each other, Yuris system proposes to address each of them as a network within each other. Three main principles guide this model (Yevdokimov, 2003)
Aggregate sustainability measure instead of sustainability indicators
Treatment of sustainable transportation as a part of a three-dimensional system economy-environment-society instead of separate evaluation of transportation impacts on economy, environment and society
System dynamics approach instead of benefit-cost analysis.
By looking at all aspects of the issue as an aggregate, a more comprehensive outlook is produced. The first principle does this by creating aggregate measures to analyze various parts of sustainability. The Genuine Progress Indicator, an overall measure of economic activity was chosen as a more complete aggregate measure than the GDP. It is integrated into a complete picture of economic activity weighed against the cost of transportation.
The second principle inserts each part of the model into a greater network of the economy in what Yuri (2003) calls a systems approach to sustainable transportation. Horizontal linkages are formed between transportation centered economy, society and environmental indicators and these are linked vertically to greater systems of economy and the environment. The linkages are dynamically integrated so that a change in either a horizontal or vertical linkage will register a result in all of the areas of sustainability (TAO, HUNG 2003).
The third principle is centered on a systems dynamics approach. It is an approach that introduces a change of system dynamics over a function of time. Instead of the linear equation seen in the optimization model a dynamic system can evolve over time becoming more adapted to the purpose for which it was built. This concept can be described as a system of mathematical equations that can essentially be boiled down an equation of this type (Yevdokimov, 2003)
Vt1 VtVt(F,P)
With Vt1 being the condition of the transportation system next period, Vt being the current transportation system and Vt being the change of the transportation system. F and P are the economic and policy variables respectively.
The aggregate constraint placed on this model to relate it to social sustainability is
GPIt1 e GPIt
With GPI being the economic indicator discussed earlier. As it is used here, GPI is an indicator not necessarily of economic sustainability but of social well-being. The equation relates issues of transportation to the betterment of society at large with the main goal being in line with sustainability, to have a continuous improvement of well-being over time (Yevdokimov, 2003).
The STS as described by Yevdokimov is a different approach to sustainable transportation then what has been classically adopted or even what is commonly used today. It places the system of transportation within a broader context of the world via vertical linkages while also intraconnecting the commonly accepted problem of transportation through horizontal linkages. Unlike optimization models it also dynamically links all of the parts of the model so that as one part of it changes all sub-parts change. The introduction of change over time, or more appropriately the constraint of an increase in well-being over time, and the use of aggregate indicators instead of individual novel indicators are an approach that is unique to this model.
It is difficult to compare the STS model to other models since it is such a different method of analysis. A few advantages and disadvantages do stand out. First of all, Yevdokimov pays lip service to the necessity of environmental concern in his paper but does not spell it out as he does economic and societal concerns. This may simply be a an oversight in the paper whereas there is still a place for it in the computational model but equations used do not seem easily applicable to anything environmentally related nor is there an aggregate indicator described as there is for other parts of the model. An introduction of a unifying aggregate measure would vastly improve this model. Linking this vertically to other levels of society, construction for instance, would be difficult since not all industries are concerned with the same parts of the environment, whereas the economy and society are more universal and easily defined.
Another issue with this model is that of specificity. Putting aside the neglect of the environments share of transportation sustainability, the model does not appear to be very specific about the indicators used. Understandably, not all indicators can be listed within a single paper but it would appear that the outcome of most the computer analyses produced by this model are simply listed as a change in GPI over time. As defined by Yevdokimov, this is a general state of well-being. It doesnt state exact values of change nor what changes can be made.
Figure SEQ Figure ARABIC 4 Changes in GPI over 2003-2025 (Yevdokimov 2003)
It would seem the STS is a good start to a great idea. It remains broad scope in order to successfully bring sustainable transportation into a much larger definition of sustainability. The vertical linkages are a powerful concept that can and should be utilized in other models. The general concept of the paper is to have a maintenance or improvement of well-being over time (Arrow 2002). By placing this a constraint on the model instead of a goal the idea is to have this as a purposeful outcome. Yevdokimovs ideas are uniquely useful to a broad range of ideas but in order to be effectively implemented they would be better utilized within a model that specifically all of the issues that are typically accepted to be associated with sustainability.
Sustainable Urban Transportation (SUTRA)
The SUTRA project is actually a consortium of other computational models. It is based on the idea that no single model can cover all of the processes necessary and the interconnections between those processes. Instead, a loosely combined set of models, set within a network of common indicators was developed within SUTRA (Fedra 2004). Although, the project is actually composed of numerous other models the indicators associated each with each model actually form a coherent network of indicators for baseline analysis, ranking and benchmarking. The tools used for this development are listed as (Fedra 2004)
Technoeconomic analysis and energy systems analysis and modeling using well-established modeling approaches, such as MARKAL
Traffic equilibrium modeling was used to evaluate alternative transportation policies, including multimodal systems, high-occupancy vehicles, park-and-ride systems, and transportation telematics and their relation to land use, technological development, socioeconomic development, and spatial and structural urban development (land-use scenarios) in general.
Emission modeling that translates the results of the transportation model
Air quality modeling is used to translate emission scenarios into ambient air quality estimates.
A fussy-rule-based system was used to estimate public health impacts and the probability of and costs of accidents.
The economic assessment, using econometric valuation methods
Finally, a discrete multicriteria method was used for a multicriteria ranking analysis
The basic method of this model is essentially that of the basic indicator based approach described in the introduction. A set of indicators is defined, a set of acceptable ranges for those indicators is established, an analysis of those indicators with the various models is conducted and those analyses are organized using a multicriteria method. At the root of the definition of the indicators is the same set of objectives as presented in the Brundtland Report (Fedra 2004), just like every other computational model. This model is unique because it uses individually constructed computational models that sometimes were not designed with the transportation industry in mind and brings them together with an overall computational method. A discussion of each major model is necessary to describe the SUTRA project further.
MARKAL is a large family of models designed as a mathematical model of the energy system of one or several regions that provides a technology-rich basis for estimating energy dynamics over a multi-period horizon (Loulou, et al 2004). In the case of the SUTRA project an adaptation of MARKAL, labeled MARKAL-lite, is used. It provides an analysis of total regional energy usage with respect to two main criteria, the type of transportation utilized (technological choice) and the emissions created by the entire system as related to areas outside of that sector (Fedra 2004). This is a complex program that has been implemented in several countries mostly to measure levels of greenhouse gas emissions and how they can be minimized by introducing alternative types of energy. Other than being an extremely useful tool in assessing the environmental impact of transportation models the MARKAL model is easily adapted for integration with the other models used within the SUTRA project, including the traffic equilibrium model, and a geographical information system.
The TERM model is used within the SUTRA superstructure as an emission modeling software. It is a comprehensive set of environmental indicators developed by the European Environmental Association (EEA) to assess emission levels across the continent (EEA Report 2010). For the SUTRA project it is integrated with the equation Eea with E being the amount of emission, e being the emission factor per unit of activity and a is the purpose of the transportation. The emission factor is changed according to efficiency of the vehicle per capita and is therefore tied into the MARKAL model of energy usage. Between the MARKAL model and the TERM model a nearly total view of the environmental impact can be observed.
Several other powerful models overlap these two models to address hidden aspects of sustainability such as hot spots formed by street canyons, transportation modeling to assess the effect multi-modal forms of transportation have on air quality and health modeling. To bring it all together a Multicriteria Decision Support System (DSS) was derived, which is further described as a discrete mulicriteria optimization approach (Fedra 2004). Essentially an algorithm is derived to determine the optimum outcome based on multiple, simultaneous criteria as set forth by the underlying models that produce raw data (Ehrgott, Gandibleaux 2002).
The SUTRA project is by far the most complicated model of all those discussed in this paper. The level of complexity has its own inherent advantages and disadvantages. The overlap of the different models provides a level of redundancy not seen in many of the other computational models. Sets of data that may be missed by a single model are outlined in the SUTRA project. Air quality indicators are exhaustively defined and analyzed using multiple models that in and of themselves are used as stand alone models for emissions standards. The TERM and MARKAL models are often used for national level determination of acceptable levels of greenhouse gas emissions (Broek et al 2009). Each addition of indicator analysis provides another level of detail that is useful in determining the overall sustainability of a region.
As defined by Tao and Hung (2003), the SUTRA project is a multi-level index model within the class of assessment indicator models. It is considered a multi-level index model because it places each model used into a set of criteria that is then weighted based on a predefined optimization level. It is unique in this regard because underneath the SUTRA umbrella are multi-dimension matrix models like the MARKAL model or other indicator models not mentioned here like AIRWAVE, another type of air quality assessment. The matrix models are typically as more complex and therefore superior to index models but in this case the organization of the project takes a proverbial step backward.
Simply adding more models to an overall decision making piece of software is not necessarily going to make a single powerful comprehensive model. Redundancy of information can be useful in catching bits of the scene that may be lost otherwise but it can just as easily create conflicting evidence as multiple data models produce multiple sets of information. Aside from the possibility of opposing sets of data, the simple level of complexity makes the SUTRA projects difficult to handle. Granted, it is bottom-heavy, which is where one would want a model to be lopsided if it had to be at all, but without a central organizing head it becomes unwieldy. The only concrete computational method bringing together this mass conglomeration of models is the DSS designed by Kurt Fedra and his colleagues, the author of the main publication describing this system. That means the entire set of indicators defined by whatever group decides to use this model must insert their criteria into the central DSS and rank those criteria according to levels of importance. This is so that the decisions made by the program may be optimized according to how the user sees fit. There is no backwards manipulation of the indicator assessment models nor does the DSS adjust the individual models to account for changes in society, economics or the environment. So the centralizing computational model can be seen in one of two ways. It can be an easy to use localized model that organizes the results of multiple other models according to what indicators are deemed important or it can be limiting because it does not intuitively change with the dynamics of its surroundings and it doesnt actually articulate its individual components.
The economic and environmental parts of the sustainability triangle are well taken care of with this model, but the social side is seemingly left out. This is the opposite problem of Yevdokimovs dynamic model. Air quality and cost-benefit analysis are extensively laid out, playing a major part in the projects overall definition of sustainability. However, the societal acceptance of any single outcome is not addressed. A concept similar to Yevdokimovs aggregate GPI measure would do well here. A complete picture of total sustainability could then be produced with actual social well-being in mind as well as the environmental and social aspect.
The SUTRA project is correct in saying that no single model is currently able to encompass all the problems with sustainability. There is no one model capable of doing this right now, so it is entirely possible that the SUTRA project may be the best method of producing a model that can bring together all the ideas of sustainability and apply it to the many different areas of society. This multi-model approach lends itself well to current culture as well. The new Iphone, with its countless apps and centralizing architecture is a similar concept that seems to be wildly successful in its implementation.
The SILENT Model
The Sustainable Infrastucture, Land-use, Environment and Transport (SILENT) model is a computational model of sustainability that actually deals with transportation as only a portion of a much larger concept of sustainability. The model follows much of the same generic process as outlined in the introduction but does so in a cohesive, planned out manner. The working definition of sustainability as defined by the designers is the long-term viability of urban living that minimizes the negative impacts of urban demography, land use, urban form, and transport on the environment, (Yigitcanlar Dur, 2010). Immediately, the direct outline and scope of the model is laid out plainly by their definition of what sustainability means to the project. Beyond the definition, SILENT follows an indicator assessment approach with a complex index analysis model at its core that ultimately outputs a graphical interface on a Geographic Information System (GIS). The explanation of the process is actually particularly easy to follow.
The SILENT model follows four steps in collecting, analyzing and ultimately outputting data. The relevant indicators are chosen first. These are decided by sustainability concerns in the area in which the model is being applied. Next, sets of parameters, definitions and categories are applied to the indicators. Once that is figured out the values are aggregated via a normalizing factor to produce a composite index. A simple index model is perceptibly used here but in fact a complex network of multivariate analysis is used to determine the composite index. Finally, once that is complete the data is output onto a GIS in grid cells that read as levels of sustainability measured from 1 to 5. The output is easy enough for nearly anyone to understand the theory.
The third step of the process is where the real power of the computational model is at. The determination of the indicators and the acceptable levels of each are set by policymakers outside of the construction of the model, but overall optimization of those indicators lies completely within the architecture of the model. The relative value of each of the indicators is determined via a Delphi study, an expert team of consultants, and several statistical methods. The values are then normalized and aggregated to provide an overall score of sustainability to each individual area studied with a regional scope. The outcome is a single calculated number on a Likert scale from 0 to 5 with 0 to 1 being low and 4 to 5 being high.
The description above is the overall computational method by which the SILENT model is built, but it is really just an aspect of a four part base system. What really drives this model is an interconnection between a conceptual base, the indicator base described above, an indexing base and a policy support base. The idea is that each base of the system informs the next. The conceptual base is at the beginning of the process and encompasses the societal ideology of sustainability. That determines the necessary indicators used for the next base, aptly named the indicator base. The type of indicators used are then indexed according to weighting factors put in place according to the Delphi study, experts and other methods. Finally, all of it falls into place in the policy support base where it guides future policy and scenario development. The loop completes itself as the future scenario then shapes what the culturally accepted meaning of sustainability is, and therefore changes the definitions in the conceptual base (Yigitcanlar Dur, 2010).
Figure 5 Structure of the SILENT model
Possibly the strongest section of the SILENT model lies in the GIS interface. Thanks to the aggregation and normalization of the composite index values drawn out by the multiple steps of the indexing base, the ultimate outcome of the model is laid out on a GIS in individually valued grid cells. The policy changing possibilities of this easy to use interface are enormous. The information is typically analyzed by those that arent necessarily capable of understanding the difficult language inherent in so many other transportation computational sustainability models. Furthermore the software can be molded into parcels of land on a physical map instead of grid cells. This way an exact sustainability factor for each land-use area can be determined and adjusted for.
The indicator based approach is essential to the simplicity and consistency of the SILENT model. It is continuously manipulated by the policy support base of the model which is actually driven by the governing bodies surrounding the region in question. On the other side of this attribute of the model is the possibility of an oversimplification of an otherwise complicated subject. One of the benefits of the SUTRA project is its redundancy in data acquisition. In an effort to be concise in the levels of indicators chosen, the designers of the SILENT model may be omitting vital pieces of information.
There are several other areas for improvement for the model as well. An expansion of the indicators is absolutely necessary. As it is laid out now, there are no direct indicators for economic status. The societal and environmental pieces of the sustainability triangle are built into the model but there is no economic feasibility allowed for in the design. An aggregate index similar to the one developed by Yevdokimov in his STS model would be best applied within the indexing base of the model as a constraint of the weights applied to other society and environment values. Also, the visual output system is extremely useful in its application to laymen but it begs the question of where to go from there Once a parcel of land is deemed unsustainable there is no guidance as to what can be done to improve that area. The multivariate analysis of the model should allow for a multimodal output of indicator values. This way, the actual situation could be compared against multiple other scenarios.
Conclusion
The words that Fedra used to describe the SUTRA project were no single model can cover the entire range of processes (2004). With each new model built to solve the problem of transportation sustainability this becomes both more and less true at the same time. The three main objectives of improving the quality of society, the economy and the environment remain the same no matter what definition is used as the conceptual basis of sustainability especially as it associated with transportation. Because of that, any future computational model must address all three pillars of the matter and it seems as though they are becoming successively more adept at doing so.
Of all the computational models compared in this paper, the SILENT model was most successful in addressing all of the issues of transportation sustainability while also remaining clear and focused in its purpose of overall sustainability analysis. TAO and HUNG explain that there is always a tension between comprehensiveness and convenience (2003). The SILENT model walks this fine line better than any of the rest compared, largely because of the end result of the analysis of data. The GIS interface is remarkably easy to understand. It is vital to the sustainability of the model itself that it be easily applied, otherwise its usefulness is lost to those that may actually be able implement changes to the culture of transportation.
The SILENT model is not without its own shortcomings. As addressed before, it does not confront socioeconomic barriers to improving the transportation infrastructure. This is a major shortfall. Perhaps a union between a systems dynamic approach of aggregate social well being and the four base system at the root of the SILENT model would be beneficial. This may not be feasible or it may produce an entirely new model that would finally encompass all of the factors seen by society as central to the concept of transportation sustainability.
In the end, no one model is currently capable of analyzing all aspects of society, the environment, and the economy. The ultimate goals of sustainability as a whole and especially that of the transportation sector are well known. The criteria for development are generally understood, it is simply the creation of a centralized, easy-to-use and comprehensive method of analyzing those criteria that is elusive to the construction of a total computational model of sustainable transportation.
Categories:
0 comments:
Post a Comment