Feasibility Study Call 1
Call Announced: February 2016
Application deadline: May 2016
Decisions to applicants: June 2016
Start/end date: June 2016 - Dec 2016
Funding available: Up to £40,000 to be divided amongst the selected proposals
The submissions will be reviewed according to the parameters stated in the call:
Match to network remit
Engagement of early career academics
Realistic objectives within the constraints of funding and time
Multi-disciplinary and innovative science
Value in enhancing research activity
Alternative funding sources
After a thorough review process by the Management Team and Steering Committee, the following were awarded funding:
Professor Alan Purvis, Durham University - £6,000
Title: Bio-Inspired self-healing strategies triggered by big-data anomalies in systems exhibiting complex behaviour to enhance their resilience.
The technically-complex scenario concerns aircraft health. Over 120x1kHz variables define the systems current state. Potential problems can be diagnosed early as the data drifts from normality. Big-data cluster analysis can detect and isolate faults triggering a restorative response. Convergent cellular automata, a class of automata which can pull the system back toward acceptable behaviour will be developed. We have previously studied the self-healing properties of convergent cellular automata (CCA, Jones) and their scalability (McWilliam).. We believe natural resilient systems use similar mechanisms i.e., multi sensor monitoring
and healing scenarios where the machine and environment need to balance before health is regained. The socio-economic scenario will look at indicators identifying a crash and examine how to embed CCA healing rules e.g., limiting credit to prevent catastrophic failure. Our closed loop systems will require extensive monitoring coupled with approaches to quantify resilience. We are grateful to NI for making available their latest hardware and software enabling trials equivalent to transatlantic flights or global trading over a decade.
Professor Jonathan Lawry, University of Bristol - £9,036
Title: Opinion Pooling in Robot Swarms
Reaching consensus across a population of autonomous agents is a fundamental requirement for decision making and planning in swarm robotics. Application domains include using large swarms of robots for pollution monitoring and treatment, search and rescue, and self-reconfiguring sensor networks. In complexity science there is an extensive theoretical literature on opinion pooling in agent-based systems. In this feasibility study we propose to test out this theory on actual robotic swarms. Experiments will be on robot swarms of up to 1000 kilobots available at the Bristol Robotic Laboratory.
The swarms’ sensitivity to learning parameters will be investigated and, in particular, their effect on convergence, quality of decision making and robustness to failure and environmental change. The study will provide insight into the real-world applicability and robustness of opinion pooling algorithms as tools for distributed decision making in actual swarms, as well as grounding current theoretical results in opinion dynamics in a physical environment where the experimental conditions can be strictly regulated. Our overarching aim is to build an understanding of how different pooling operators influence the high-level decision making capabilities of the swarm both in terms of the level of consensus reached between agents and the quality of the decisions made.Lawr
Professor :Lawry recently had a paper accepted at the ROS Robotics conference on the workRoboti funded by the ENCORE Feasibility Study. This is one of the leading Robotics conferences.
Dr Thilo Gross, University of Bristol - £9,423
Title: SKYLINE: Finding the variables of the city
In real world complex systems a central challenge is to find the variables that can be modelled meaningfully. This is particularly true for complex engineered systems such as cities. The development of a city at a macroscopic level is the result of a myriad of individual decisions by humans. Such a system can modelled easily on a macroscopic scale, but then detail existing on the microscopic scale of the individual is lost. By contrast capturing all of the microscale interactions and choices of humans results in models that are at best unwieldy and often infeasible. But, in many complex systems a mesoscale exists at which a
detailed but manageable modelling is possible. While good mesoscale variables may be non-obvious, they can be constructed mathematically. For a city this may for instance reveal a characterization at street level, where the physical and social properties of every street segment are described by a small set of variables. Here, we propose to combine several openly available datasets for the city of Bristol and perform preliminary mathematical analysis that should reveal whether a mesoscale modelling of the city is possible and identify suitable variables on the mesoscale.
Find out more about Dr Gross's work and his latest ENCORE Workshop on Maths & the city.
Dr Chrisos Ellinas, University of Bristol - £10,000
Title: Developing methods for project resilience against failure cascades
Projects are the de-facto medium for organisational activity, with project performance being intrinsically linked to economic prosperity, at organisational, industrial and national level. Yet their complex nature is increasingly challenging our capacity to successfully deliver them, often resulting in wasted resources and missed opportunities. Recent work suggests that project complexity is at the core of this challenge, emphasizing the inability of traditional project management tools to account for the resulting effects of this complexity (e.g. non-linearity). Fuelled by these effects, recent work has highlighted the capacity of projects to sustain
failure cascades, in the form of task failures, where the failure of a single task can affect a significant portion of the project. Developing means for minimising the impact of these failures can contribute to project resilience, and in effect improve project performance. As such, this proposal will develop a mitigation strategy for containing the impact of these failures, along with proving the means for recovery. Importantly, this strategy will satisfy a given set of restrictions that arise due to the distinct nature of projects. In terms of data, this proposal utilises readily-available datasets, in the form of Gantt charts, strengthening its operational focus and practical utility.
This proposal resulted in a successful 2-year fellowship application by the Dr Ellinas. The Fellowship is funded by EPSRC and to will be hosted at the University of Bristol (UK). The fellowship project officially launches on 01/04/2017 and is entitled “Application of Complex Networks to Project Robustness and Resilience”.
A full report of the Feasibility Study can be found here.
Dr Neil Carhart, University of Bristol, £3000
Title: Do engineering standards support or undermine the creation of resilient systems?
A one-day workshop took place on 24th November , convening the authors and users of engineering standards to explore the role they play in creating particular causal structures which then influence the resilience of the systems they are used to design. This workshop utilised Causal Loop Group Model Building
based on formal problem structuring methods, building on a previous workshop involving the Rail Safety and Standards Board (RSSB) and the Office for Nuclear Regulation (ONR). Participants were presented with simple models based on historical incidents of systemic fragility and failure as a basis to guide discussions around the potential role played by engineering design standards. This will result in a set of models, verified by expert analysis, hypothesising the causal mechanisms through which standards may undermine resilience.
Share the NEWS