The recent advances in pervasive technologies enable construction of large-scale socio-technical systems which tightly interweave humans and their social structures with technology. The overall goal of Allow Ensembles is to develop a new design principle and establish a new foundational framework for collective adaptive systems (CAS) based on the concept of cell ensembles.
In Allow Ensembles each cell carries a genome, which is modelled as a declarative flow describing the possible behaviour of the cell. When a new goal is introduced, the system autonomously configures an ensemble of cells to accomplish the goal given the current context and a utility associated with the ensemble. Declarative flows specify the activities that may be executed as a part of the flow execution and a set of constraints that the execution must satisfy. For example, a constraint may define the order of two given activities or whether or not an activity is performed depending on the flow’s current context. In its most general form, a declarative flow just defines a set of activities, which may be performed in any order.
Allow Ensembles will develop new models, theories and algorithms that can:
- Autonomously form large-scale collective adaptive units (ensembles) that can flexibly satisfy arbitrary goals in the real world environments.
- Constantly improve the utility of ensembles by adapting the individual cells within an ensemble such that the overall system learns to get better at accomplishing goals.
- Make ensembles robust such that they can survive a wide range of hardware and software failures and still accomplish a given goal with a high utility.
- Evolve ensembles in order to promote beneficial emergent properties and suppress detrimental emergent properties.
- Make ensembles secure and protect sensitive data by evolving security policies in unison with ensemble evolution.
The validity of the the framework will be investigated using two scenarios that are become increasingly relevant on the European scale: integrated urban mobility and smart production chains.