EME Taster October Department of Computer Science EME: Emergence Module Susan Stepney Fiona Polack
EME Taster October Department of Computer Science What is an emergent system? Many simple elements, no central control Overall behaviour that is not an obvious consequence of behaviour of elements –Construction by insects –Crowd or flock behaviour by people, birds, insects –Plants grown using simple growth algorithms –Animals developed from simple cells
EME Taster October Department of Computer Science Why study emergent systems? Emergent systems cannot be understood by direct extension of conventional systems ideas An emergent system is a system of systems –Emergent properties at system level –Components at system level* Emergence requires a change in scale –Small systems combine to larger-scale effect * This will not make sense if you are colour-blind or reading in b/w!
EME Taster October Department of Computer Science How can an emergent system be engineered? Through studying lots of emergent systems –We’ll cover various sorts in EME By considering levels and the environment –Polack and Stepney work on architectural aspects of emergent system development Through simulation and real-world observation –What needs to be modelled to faithfully simulate emergence? Paying attention to dependability assurance…
EME Taster October Department of Computer Science What will we look at? L-systems Measures and uses of complexity Cellular Automata & variants Information theoretic views of emergence and self- organisation Some definitions of emergence Engineering emergent systems Examples of emergence in natural systems Nano-construction
EME Taster October Department of Computer Science CA: demonstrate complex behaviour from simple rules CAs Time 1Time Transition Table rules... Game Of Life CAs wonders/life/life.html Langton’s self-replicating loops /sdsr/java/
EME Taster October Department of Computer Science What will we look at? Demonstrations that complex behaviour emerges from multiple instances of simple systems: L-systems –Aristid Lindenmeyer & Przemyslaw Prusinkiewicz –biological metaphor: plant branching growth eg. a bud becomes a twig, a twig gains buds, …
EME Taster October Department of Computer Science What will we look at? Demonstrations that complex behaviour emerges from multiple instances of simple systems: L-systems –Aristid Lindenmeyer & Przemyslaw Prusinkiewicz –biological metaphor: plant branching growth eg. a bud becomes a twig, a twig gains buds, … –a production replaces a parent module with a child module pn: symbol rewrite_string “bud” “twig”p1 : B T “twig” “twig” “bud” “bud”p2 : T TBB
EME Taster October Department of Computer Science What will we look at? Demonstrations that complex behaviour emerges from multiple instances of simple systems: L-systems –Aristid Lindenmeyer & Przemyslaw Prusinkiewicz –biological metaphor: plant branching growth eg. a bud becomes a twig, a twig gains buds, … [Prusinkiewicz & Lindenmeyer, fig 1.24a, c, d, 1.10, 1.24f, 1.8]
EME Taster October Department of Computer Science What will we look at? Demonstrations that complex behaviour emerges from multiple instances of simple systems … and various other artificial complex systems
EME Taster October Department of Computer Science What will we look at? Demonstrations that complex behaviour emerges from multiple instances of simple systems: Nanotechnology –“molecular” nanotech little things that make big things –Nanobot constructors and replicators Design for a fine-motion controller for molecular assembly atoms, <1nm
EME Taster October Department of Computer Science Engineering Not much work has been done on engineering –Most artificial emergent systems are hacked Conway’s Game of Life CA was discovered by playing with patterns on a GO board Proposals for real-world nano-scale assemblers and disassemblers make engineering urgent We’ll look at –Identifying and measuring emergence –Layers in emergent systems –Towards assurance-directed development frameworks