-In this study, we elicit and specify a goal model for the software engineer capturing, detailing and analyzing at the the recycled water system of Abu Dhabi Emirate, along with the design stage the precise knowledge and decision measures that constraints and strategies defined within the system boundary. will guide self-adaptation at run-time. Its technique brings aTphperoparcehlimbiynamryodreesliunlgts aanrde esvimaluulaatteidngustihneg gSoyasltsemansdDyrenqaumirices- the high level requirements, in the form of goal-models, to ments defined using Tropos for Adaptive Systems (Tropos4AS) operation time, allowing the system to reach their satisfaction, method. System Dynamics methodology allows investigation of to reflect upon them and to guide its behavior according to the interactions between goals and the evaluation of the results. them [4]. This paper demonstrates preliminary evaluation of the System Having a rigorous model helps generating accurate forecasts DMyonraemoviecrs, atbhielitlyinikninvgaliodfatSiynsgtermequDiryenma menitcss eanngdineTerroipnogsam4oAdseliss. thus facilitating the effective planning and decisions making. found promising since it seems it can improve the collaboration Therefore, this study aims to use a Tropos4AS to specify goal between goals of different stakeholders for better planning, models, with the assistance of System Dynamics approach, by considering the prioritized goals resulting from the hybrid that characterizes the recycled water system of Abu Dhabi analysis. Emirate. The paper presents related work, research method, posI4nAdeSx, sTyesrtemms-dgyonaalm-oircise,nwteadterresqeucuirreimtyents engineering, Tro- models development, results and evaluation.
Significant research has been devoted to the water scarcity
issue due to its importance and the impact that it could
create in hindering the sustainable development [
In fact, managing water resources efficiently is a necessity, and the potential of recycled water resource to improve the water system could be tremendous. In this context, many methods could be used to develop this complex system such as optimization, system dynamics, system modeling language, etc. However, these methods should start by precisely specifying the goals and requirements as a critical step towards taking correct decisions where any mistake could lead to huge losses. In order to develop an efficient system, we need to have a clear understanding of its main goal and its operation. Basically, this involves a complex process of identifying, recognizing and deciding on the system’s goals, requirements and functionality.
Requirement Engineering (RE) consists of analyzing and
outlining the set of activities that are essential for structuring
and constructing a system where the functionality and the
outcome meet the stakeholders requirements [
GORE methods from RE provide a systematic approach for
extracting, evaluating, analyzing, modeling and determining
software requirements. These approaches are built based on
certain fundamental concepts and principles, they focus on the
needs of the stakeholders and express their requests in a natural
way, concentrating on their requirements and objectives, rather
than how to meet them [
Numerous GORE languages and approaches are available
in literature (e.g., Keep All Objectives Satisfied (KAOS),
i*, Security Quality RE (SQUARE), Tropos, etc.) KAOS
is a GORE approach initially applied by systems analysts
to satisfy the requirements through the achievement of the
goals [
The Tropos approach, is considered as an agent-oriented
software engineering (AOSE) approach that covers the entire
software or system development process. This methodology is
based on two main ideas: firstly, the concept of agent and all
the related notions, such as goals and plans that are used in all
stages of the system development, from the early investigations
and analysis down to the actual application. Secondly, Tropos
concept covers the very early stages of requirements analysis,
consequently allowing for a profound understanding of the
environment where the software will operate [
In his study, Morandini developed a framework for
engineering adaptive systems called Tropos4AS. Its framework
provides analysts with modeling features to enhance the
requirements analysis on the precise knowledge and decision
criteria that an adaptive system needs to autonomously adjust
to dynamic changes. It also contains a planning to software
agents for prototyping and performing requirements
simulation. Accordingly, it helps analysts in their requirements
validation and refinement phase. Tropos4AS framework is
based on three key modeling features aiming simultaneously
to capture the specific requirements knowledge from, and to
decrease the gap between requirements-time abstractions and a
run-time representation of requirements for adaptation as well.
These three key modeling aspects are [
System Dynamics approach was developed by Jay Forrester
at Massachusetts Institute of Technology in 1950 [
Basically, the water supply sources in the Emirate of Abu
Dhabi have not witnessed a noticeable change since 2009
[
Recycled water is managed by Abu Dhabi Sewerage
Services Company (ADSSC) starting from wastewater collection
to treatment down to the disposal of bio-solids and excess
recycled water. Unfortunately, only 26% of water used by
domestics returns to sewer due to poor demand side
management of the public, such as irresponsible outdoor water
usage including but not limited to car washing, irrigation,
pools filling, etc. Moreover, about 25% of the desalinated
water is lost from the system due to inefficient transmission
and distribution or illegal consumption. However, some efforts
are placed to face such losses and decrease them down to an
average of 17.5% by 2030 as announced by the Regulation
and Supervision Bureau in Abu Dhabi [
On the other hand, only 52% of recycled water is used
while the rest 48% is discharged into the sea or in the open
desert because of insufficient transmission and distribution
capacity [
Copyright © 2017 for the individual papers by the papers' authors. Copying permitted for private and academic purposes. This volume is published and copyrighted by its editors.
Currently, the Enhanced level (beyond Tertiary) treatment
is considered only in a pilot project at Al Nahda farms for the
purpose of irrigation relieving the use of groundwater [
While the fourth level treatment technology is one of the
future options, as announced by EAD and ADFCA [
This technology will be able to treat industrial wastewater for
the same sector usage [
In this section, an overview of the analytical assessment
approach followed in this study is explained. The investigation
of the characteristics of any system qualitatively is considered
fundamental. It allows for understanding of the complexity of
the system and its behavior. Thus, we first investigate the status
of the recycled water supply system in the Emirate of Abu
Dhabi to understand its state and dynamic behavior, along with
the demand side. The objective of this step is to get an insight
about the system actors who have an influence on system
complexity due to requirements and goals prioritization. More
specifically, we model the recycled water current status and
strategies placed (or aimed to be in place) towards achieving
the government goal of maximizing the recycled water usage
and enhancing the system accordingly [
By focusing on Abu Dhabi Emirate case study, we aim to
evaluate the capability of Tropos4AS approach in
characterizing complex systems. With the assistance of System Dynamics
approach, we aim to evaluate the goal models developed by
Tropos4AS. Such step will enable assessing and updating
developed goal models if change in requirements are needed
towards making the system more rigorous, through testing
their potential empirically in System Dynamics. The research
method followed in this study is as follows:
1) Data collection: the data used in this study is collected
from literature, including governmental reports,
brainstorming, observation, and discussion with an expert
from the field of water resources management (EAD).
Few assumptions are made:
• Public Acceptance Rate is considered to be 42.8%
1 [
The Tropos4AS modeling methodology has been defined considering the key properties that need to be engineered into an adaptive system. This methodology helps the system to adapt autonomously, by using a combination of current capabilities and conditions at run-time. The assumed stakeholder in our case is Abu Dhabi Government and the actor is ADSSC, they are both the decision makers of our project. Once the stakeholders and actors have been assigned, the second step is to define the goals. Our main goal is to enhance the recycled water system in Abu Dhabi Emirate. This can be achieved by ensuring the satisfaction of two goals: increasing the recycled water usage from one side and improving the efficiency of the system from the other side.
It is essential to include in our system the main goals to reach as well as the sub-goals, to specify what effects the environment has on the system at run-time in addition to the 2Variable names between brackets are as used in System Dynamics Model in Vensim PLE potential errors that lead to the failure of our system, and the possibility and plans to monitor these errors to get back our system to the right track towards achieving the main goal. As aforementioned, the Tropos4AS modeling methodology concepts are extracted from i* methodology, and its development procedure is based on the agent-oriented software engineering methodology Tropos which focuses mainly on the actors representing the system-to-be. Consequently, Tropos4AS allows modeling requirements taking into consideration the behavior of the system to satisfy all its goals. The model includes three main components: • An extended goal model with the corresponding run-time satisfaction conditions. • An environment model indicating the critical key elements in the operational phase of our system that affect the goal satisfaction. • A failure model pointing out any potential errors that may lead to the failure of our system consequently preventing the system from achieving its goals and the inclusion of the necessary recovery plans.
Extended Goal Model: Tropos4AS methodology includes information about the dynamics of goal satisfaction in goal models, which is a characteristic of this approach to allow monitoring the goal satisfaction as well as the decisionmaking in an adaptive system at run-time. The extended model includes the main goals as well as the sub-goals in addition to the plans required and necessary to apply for the satisfaction of these goals. Figure 2 shows the extended goal model of our case study.
In order to fulfill the main goal, which is to enhance the recycled water system in Abu Dhabi Emirate, we need to ensure the satisfaction of the two major components in the model which are the increase of the recycled water usage and the enhancement of the system efficiency. The sub-goals related to the satisfaction of these goals are included in the model. First, increasing the collection, by improving the collection of the recycled water we can increase the recycled water usage and consequently achieve our main goal. The plan necessary for the satisfaction of this goal is to enhance the return to sewer rate as well as the improvement of the transmission system. Second, increasing the transmission capacity, where this subgoal is one of the essential ones to increase the recycled water usage as well. The necessary plans are to develop projects to improve the plants’ capacities and to enhance efficiency of the transmission in the system. Third, the public acceptance which is another important aspect affecting the usage of the recycled water. To achieve that goal, we need to treat the recycled water to the enhanced level, after the tertiary one, by using advanced technologies. Using such an advanced treatment should lead to the population having more trust in the purity and safety of this water.
Finally, it is necessary to develop more enhanced technologies used for the recycled water process. By doing so, we are ensuring the satisfaction of both main goals, increasing the recycled water usage and the improvement of the system efficiency. The essential plans needed are first to enhance the wastewater treatment process and then the privatization of the recycled water related project that is considered as a long-term plan.
Environment Model: Adaptivity’s main concept is how the behavior of a system is modified whenever any kind of changes happen in its environment in order to keep its functionality as required. The environment model points out the dependencies among the agents and stakeholder goals and its intentional and non-intentional environment at the run-time stage. Tropos4AS methodology adds the concept of conditions and context in order to outline the run-time goal satisfaction procedure in response to any environmental changes. In this regard, precise conditions are connected to goals, showing the constraint or the context necessary in the environment for the achievement of this specific goal at run-time.
In our case study, and as represented in Figure 3, the first condition is related to the collection control. We need to maintain a collection of not less than a 26% of the wastewater, which is the current amount in Abu Dhabi Emirate. Failing to do so will affect the amount of resulting recycled water produced, reducing the amount of recycled water used, affecting therefore the goal to increase the recycled water usage. The second condition to apply is associated to the usage control. The amount of the treated wastewater discharged into the sea should not exceed 48%, the current amount in Abu Dhabi. In case this volume is surpassed, the usage control will be affected and accordingly the efficiency of our system will decrease affecting directly the main goal and preventing the enhancement of the recycled water system. The third condition to apply is to sustain the current transmission capacity, and not going below 52%, which is the current amount transmitted in Abu Dhabi Emirate. Going below this limit may have negative consequences on both the recycled water usage and the efficiency of our system thus preventing the satisfaction of our main goal to enhance the recycled water system in Abu Dhabi Emirate. Finally, it is important to include the public awareness and acceptability of the recycled water. Approximately 42% of Abu Dhabi inhabitants are assumed aware of the importance of the recycled water, consequently they accept and encourage the usage. Agitating the treated wastewater system negatively affects the trust of people in the system and in the purity of the recycled water, thus decreasing its usage and preventing the satisfaction of the main goal in our model.
Failure Model: Another important feature of the Tropos4AS modeling methodology is that it introduces a failure model. Its main objective is to provide means for capturing exceptional circumstances and probable undesirable activities at the runtime stage. These activities are reflected in the model as errors leading to failures of the system. They are mainly a result of any uncompleted requirement, or a consequence of circumstances that were not taken into consideration at the requirements analysis stage or the design-time stage. The failure model helps anticipate any error of failure at the runtime phase by adding the recovery activities known as plans.
Through the following Figure 4, we explain the failure model of our case study that includes potential errors and the failures resulting from these errors, and the necessary recovery activities or required plans. First, cutting the transmission and a plant shutdown are both potential errors that can happen to our system at run-time and lead to a failure in the operation, thus leading to a disruption in the supply of the recycled water affecting directly the recycled water usage which will decrease subsequently. In order to deal with this failure, many recovery activities could be taken, such as transmitting recycled water via tankers. This action can be considered as a short-term plan action guaranteed immediately. Another short-term plan is to follow the safe disposal strategies of wastewater into the sea or desert. Additional recovery activity is the maintenance of the damages and the repair of the disruption, both of which may be considered either long-term or short-term depending on the size of the damages and the time needed for the restoration. Another set of errors presented in our model is whenever a leakage in the transmission happens or if the plants are overloaded, these errors cause a drop in the efficiency in our system and highly affect our main goal to enhance the recycled water system in Abu Dhabi Emirate. As restoration plans, we first propose to control the leakages and to provide the necessary maintenance and reparation. Furthermore, recovery activities are to propose projects to enhance the transmission infrastructure and to increase the recycled water treatment plants capacity.
Linking between the extended goal model, environmental model as well as failure model built using Tropos4AS generates a comprehensive full model for the recycled water system in Abu Dhabi.
While this is the work in progress, and we are currently integrating the results of Tropos4AS with the System Dynamics model, we have already produced some results that show the gaps in recycled water supply resulting from running the System Dynamics model over the agreed scenarios, as presented in Figure 5. So far, the scenario 2 seems to be the best having the minimum shortfall amount, indicating that the RTS rise has a great impact on reducing the supply gap. Maximizing both strategies results in the second minimum supply gap where the system could satisfy almost 33% of its demand for Tertiary recycled water and 43% of its demand for Enhanced recycled water. The BAU scenario is found better than further enhancing Tertiary recycled water transmission capacity, thus saving cost.
In this paper, we have shown the preliminary results of system modeling using Tropos4AS and System Dynamics through a case study of Abu Dhabi Emirate recycled water system. We are currently simulating the goal models and short-term plans defined in Tropos4AS using Vensim PLE, and running four different scenarios including BAU. The full results of this study will be based on the identified system boundary, modeling setup and assumptions.
In the previous sections, we have presented the applicability of Tropos4AS at the recycled water system in Abu Dhabi Emirate. Linking between System Dynamics and Tropos4AS seems promising so far. The System Dynamics method is helping us evaluate the goals built in Tropos4AS, where maximizing the wastewater collection has more impact on system improvement than maximizing transmission capacity under the given modeling setup. Thus, so far it seems that investing in maximizing the return to sewer amounts has a higher potential and could further enhance the recycled water system in Abu Dhabi Emirate. Adopting enhanced treatment technologies to treat the amount that used to be discharged has a potential too in covering a good portion of the agricultural demand, and substituting the use of depleting groundwater resources. Public awareness is indeed important for system enhancement although no effect has been noticed in our analysis so far. We believe this is due to a lower amount of supply available than impacted demand.
By investigating the literature, we could not find any application on water systems in general and in recycled water system in particular using either goal-based or agent-based modeling methods. This is most likely due to the engineering discipline distance. In our study, we applied Tropos4AS procedure to the recycled water system in Abu Dhabi Emirate, using real data and hopefully providing tangible results by the time the study is fully completed. Furthermore, this study contributes to the knowledge of stakeholders impact on recycled water system, and highlights the potential and possible risk of the current plans or decisions, thus accordingly helping in improving the system towards making it more resilient and vulnerable.
Combining Tropos4AS and System Dynamics methods have a potential to be a valuable addition to the RE research. Linking between both approaches and using the resulted unified validated model will support robust scenarios testing, thus investigating the power of potential strategies in future as well as improving the decision making. This study is the first of its kind since RE helps defining requirements for System Dynamics models but the opposite relation has never been investigated. So far, we can see both methods complementing each other effectively.
An important limitation we faced in our study is meeting the stakeholders and actors, which are in our case representatives from Abu Dhabi Government and ADSSC, to perform a deeper investigation of the interaction between system actors and goals. This restriction is preventing us from having their comments, plans, and other feedback on different conditions, errors and failures that our system may face at run-time. Having data about system design and operation could have better impact on system improvement if available. Furthermore, the study has not so far considered seasonality in demand and production capacity planning due to the limited scope in addition to the time constraint.
This paper has reported the positive preliminary results in our work on integrating Tropos4AS and System Dynamics method in order to produce comprehensive goal models that can be quantified, simulated, and used in decision making in water security domain. Given our current progress, we should be able to present full results during the workshop.
Besides the potential of enlarging the scope of this study and making it more detailed, the boundary could also be improved by integrating the other water resources in Abu Dhabi Emirate and use the resulting model as an integrated water resources management tool. Other potential strategies that could be further analyzed are: water treatment plants capacity improvement, controlling leakage from water system, analyzing the effect of water price on return to sewer rate, etc. Furthermore, the future work should also monitor and analyze the correctness of the model during the system operation. This should be done through the implementation and collection of data from the implemented system and comparison with the model. Evaluating the results could be better visualized using multi-criteria decision analysis from economic, social and environmental perspective. Finally, introducing the uncertainties such as seasonality and demand growth rate will make the analysis more representative.