||Impact Modelling in Complex Systems
Impact Modelling in Complex Systems
Prof. Mark Price
Dr. Juliana Early
Dr. Patrick Connolly
Mr. Gary Davies
The complex system that is focused on for this project is the air traffic management system. This is based on having a look at the effects to the whole system when a certain situation arises such as adverse weather conditions or a runway closure, these situations typically cannot be predicted by only examining individual behaviour alone. The model will show what happens is the effects of a situation ripple across the whole system.
The function of an impact analysis is to identify the element relationships and interdependencies as well as analysing the impact of one element on any or all of the elements within the system architecture. Impact can therefore be defined as a measure of how disrupted an element becomes due to a deviation in another elements operations. There are two types of impact to be looked at, hierarchical impact and relational impact. Hierarchical impact is measured at each level of the operational system hierarchy whereas relational impact is the impact of associated elements which are of relevance to system stakeholders. Although impact is a qualitative approach, it is converted to a user-defined quantitative measure to assist in the computational process.
Agent based modelling is used as a sophisticated way of solving problems. Agents are set up instead of setting up equations. These agents act as individual elements within a system with each agent being given a certain set of behaviours. The agents can then react in a certain way to a situation. A simulation of these individual agents acting together can predict the result of a situation which may arise. While agent-based simulations cannot include every aspect of system behaviour, it provides an interesting insight while being a quick and cost-effective form of analysis.
Paper: Understanding Aspects of Complex Systems through an Impact Analysis
G. Davies, PhD Research Student, School of Mechanical and Aerospace Engineering, AIAA Student Member
P. Connally, Post Doctoral Researcher, School of Mechanical and Aerospace Engineering
M. Price, Professor School of Mechanical and Aerospace Engineering, Senior Member
J. Early, Lecturer, School of Mechanical and Aerospace Engineering, Senior Member
The creation, development and modification of complex systems is becoming increasingly commonplace, but it remains a costly and time consuming activity to predict the behavior of such systems due to their data intensive nature. Moreover, such systems are often dynamic in nature and may not always lend themselves to quantitative behavioral analysis. The identification of interdependencies, and the consequences to the overall system state due to data transfer across these interdependencies, is key to understanding such system behavior. However, such information is normally acquired via a secondary analysis of behavioral or physical models. Such analysis of interdependencies seeks to identify when changes in one system parameter or element impacts or changes the behavior of another. With the rapidly increasing size and complexity of real world systems, even obtaining the behavioral models is difficult. Hence there is a major challenge in identifying such impacts without needing the complete physical/behavioral model. This work presents an impact analysis method as a qualitative approach aimed at measuring the deviation of an element from its ‘normal’ operational mode resulting from changes to the state of another element. The paper introduces the concept applied to an ATM system developed within an agent-based modeling environment. The results from a test case demonstrate the application of impact-based analysis techniques, and the benefits of using these tools in complex architectures to support decision making. The developed toolsets can be used to further understand the behavior of the complex systems, and the impact of change.