Gwybodaeth Modiwlau

This module contains a description of adaptive behaviour in terms of (i) systems that changes over time (behaviour), and (ii) change of a system's behaviour with respect to results of the interaction between environment and system (adaptation). It introduces the processes of adaptation, both on individual/population level, different time scales, and indirectly via changing the environment. It examines adaptive behaviour in biological systems (incl. ecosystems), individual development, agents and interactions, groups, societies, economies, etc.

The module explores mechanisms of adaptive behaviour, including: centralised vs. decentralised organisation principles, emergent phenomena, self-organization as mechanisms of adaptation and behaviour.

Finally, the module uses robot examples as tool to outline adaptive behaviour as a multi-objective adaptation process. It analyses systems in which non-linear interaction, positive feedback, noise are acting as constructive elements.

1. Introduction [3hrs]

Key concepts, Aims and objectives; Introduction of the context
used in this module (the problem of optimization);
Dynamical systems theory, basics.

2. Bio-Inspired Adaptive Systems (1) [5 hrs]

Structure and Process metaphors
Ideas drawn from animal anatomy and processes,
Computational modelling of Brain and neural systems,
Artificial Immune systems and Endocrine Systems.
The brain as a dynamical system.

3. Bio-Inspired Adaptive Systems (2) [5 hrs]

Evolutionary metaphors, Basic ideas, hill-climbing
GA for bit string representations, ES for real number representation
and self-optimisation, GP, designing algorithms for real world
problems including multi-objective functions and dynamic functions,
case study: evolutionary robotics.

5. Bio-Inspired Adaptive Systems (3) [4 hrs]

Developmental metaphors
Development as evolution of the individual, staged growth,
constraint functions, algorithmic approach, examples from
Epigenetic-robotics.

6. Adaptation from swarms and colonies [5 hrs]

Swarms: concepts, flocking behaviour, communication and control,
simulations; stigmergy, synchronisation (fireflies); Ant colonies / ACO
(ant colony optimization): motivation, implementation and applications
for NP-hard problems; concepts, search algorithms; Swarm-robotics.