#	Pagina
attuale pagina	/open-h2020/projects/213193/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - EVOAtm (Evolutionary ATM. A modelling framework to assess the impact of ATM evolutions)

Teaser

Summary

What is the problem/issue being addressed?
The research topic addressed by the project is to develop a modelling framework to perform the following analysis:
- Change impact assessment
- Support to design
- Support to strategic thinking

The three scientific pillars of the modelling framework are:
- Agent-based modelling
- Evolutionary computing
- Statistical analysis

Why is it important for society?
EVOAtm provides suitable means to support decision-making for the change design
The complexity of multi-layer ATM system makes the impact of any change to it not easily predictable starting from the micro-level behaviors of the single agents (or components). An agent-based simulation may be used in order to forecast the impact of possible changes to ATM system and to identify the optimal tuning of the change design. At the earlier stages of a change the agent-based modelling joined with evolutionary computing optimization enables a deep understanding of the emergent (i.e., unpredictable) aspects of the change, in terms of any performance variation and in terms of the possible best combination of parametric characteristics of the change. Such quick evaluation may reduce the effort spent in man-in-the-loop simulations, usually performed at the end of change life cycle.

EVOAtm provides suitable means to support strategic thinking.
Sensitivity analysis allows to identify parameters for which small variations mostly impact the model\'s simulation output and also the effect at local level. This identification may aid to establish which are the most critical individual (i.e., agent) or global (i.e., system) parameters and also highlight the interdependencies among the different agents of the system.

What are the overall objectives?

The project main goal is to demonstrate the effectiveness of the developed concept of modelling framework in being the basis of an engineered tool. Such tool shall enable the identified analyses to be performed in the earlier stages of the change life cycle at different level of abstractions.
In details the project has to accomplish the following objectives:
1) To define a suitable modelling paradigm to represent the selected ATM section in terms of the services they provide, their interdependencies
2) To setup the proper performance indicators at system level (the modelled part) and at local level in order to assess the impact of the change
3)To apply evolutionary computing in order to tune the design of new/modified ATM components accomplishing the desired performances
4) To define the proper paradigm to:
- perform change impact assessment
- support the strategic thinking intended as the decision process to assess which part of the modelled ATM system would need of a modification
5) To provide methodological guidelines in order to extend the proposed framework to a generic section of the ATM system

To demonstrate the effectiveness of the designed approach the project provides a demonstrator implementing a specific part of ATM system - the flight Execution phase in both Direct Routing and Free Routing scenario.
EVOAtm approach works using a baseline model (the system before the change), against which the changed model (the system after the change) is evaluated. In the demonstrator the baseline model is represented by the Direct Routing (the reference scenario). The changed scenario is represented by the Free Routing .
Scenario assumptions are those made in SESAR1 solution 32 and 33 .

Work performed

The main results achieved in the reported period are the followings:
Result1: A clear picture of the positioning of the EVOAtm with respect the sibling projects (PACAS, MAREA, INTUIT , APACHE and ELSA) has been addressed and thus the related level of innovation
Result2: EVOAtm formalism for specification and modelling has been defined
Result3: The project performance framework has been set up
Result4: The human aspects to detail the human actors behaviour have been almost defined
Result5: Approach to sensitivity analysis has been defined
Result6: A draft of the evolutionary computing algorithm has been implemented
Result7: The environment for the prototyping has been chosen and some preliminary part of the Reference Scenario have been implemented
Result8: A fine tuning of the project life cycle has been achieved: the waterfall life cycle has been replaced by a more efficient fast-prototyping approach
Result9: Publications and Project Information have been produced for the dissemination
Result10: The exploitation plan has been put in place

Final results

EvoATM has focused its initial effort on Sibling projects investigating them from different perspectives:
- the possible re-use of performance metrics,
- sibling approaches to change impact assessment,
- the approach to modelling and possible re-use of simulators.
Such study has allowed a clear picture of its positioning with respect such projects clarifying the level of innovation.
Where in previous projects ABM is used to model specific aspects at different level of abstractions no specific methodology is provided. None aims to define a platform enabling an exhaustive socio-technical representation of the system of interest including the human behavior modelling. Optimization is usually applier to that of trajectory whereas in the project is used for more generic problems.

The expected results until the end of the project are:
- The fast prototyping of the changed scenario
-The integration with evolutionary computing
- The identification of demonstration cases
- The refinement of the metrics