l ecture 11 c ellular a utomata 1 d iscrete t ime d
play

L ECTURE 11: C ELLULAR A UTOMATA 1 / D ISCRETE -T IME D YNAMICAL S - PowerPoint PPT Presentation

15-382 C OLLECTIVE I NTELLIGENCE S18 L ECTURE 11: C ELLULAR A UTOMATA 1 / D ISCRETE -T IME D YNAMICAL S YSTEMS 3 I NSTRUCTOR : G IANNI A. D I C ARO S TATE VARIABLES SPATIALLY BOUNDED ON L ATTICES : C ELLULAR A UTOMATA State variable


  1. 15-382 C OLLECTIVE I NTELLIGENCE – S18 L ECTURE 11: C ELLULAR A UTOMATA 1 / D ISCRETE -T IME D YNAMICAL S YSTEMS 3 I NSTRUCTOR : G IANNI A. D I C ARO

  2. S TATE VARIABLES SPATIALLY BOUNDED ON L ATTICES : C ELLULAR A UTOMATA  State variable ↔ State of a Spatial location / Cell  Cell in a 1D, or 2D, or 3D Lattice Example of selected neighborhood of 𝑦 𝑗 , represented by the set { 𝑦 𝑗−1 , 𝑦 𝑗+1 } 𝑦 2 𝑦 3 𝑦 𝑙 𝑦 1 𝑦 𝑗−1 𝑦 𝑗 𝑦 𝑗+1 1D Example of selected neighborhood of 𝑦 𝑙+𝑗 , represented by the set { 𝑦 𝑙+𝑗−1 , 𝑦 𝑙+𝑗+1 , 𝑦 𝑗 , 𝑦 𝑗+1 , 𝑦 𝑗−1 , 𝑦 2𝑙+𝑗 , 𝑦 2𝑙+𝑗+1 , 𝑦 2𝑙+𝑗−1 } 𝑦 2 𝑦 3 𝑦 1 𝑦 𝑙 𝑦 𝑙+1 𝑦 𝑙+2 𝑦 𝑙+3 𝑦 𝑙+𝑗 𝑦 2𝑙 2D 𝑦 𝑙+𝑛 2

  3. CA S ARE L ATTICE MODELS  Regular 𝒐 -dimensional discretization of a continuum  E.g., an 𝑜 -dimensional grid  Periodic (toroidal) or non periodic structure  More abstract definition: Regular tiling of a space by a primitive cell Bethe lattice, ∞ -connected cycle-free graph where each node is connected to 𝑨 neighbors, where 𝑨 is called the coordination number 3

  4. C ELLULAR A UTOMATA 1D 1D Lattice 𝑦 2 𝑦 3 𝑦 𝑙 𝑦 1 𝑦 𝑗−1 𝑦 𝑗 𝑦 𝑗+1 1 1 , 𝑦 𝑜 2 ) 𝑦 𝑜+1 = 𝑔 1 (𝑦 𝑜 1 1 , 𝑦 𝑜 2 , … . 𝑦 𝑜 𝑙 ) 𝑦 𝑜+1 = 𝑔 1 (𝑦 𝑜 2 1 , 𝑦 𝑜 2 , 𝑦 𝑜 3 ) 𝑦 𝑜+1 = 𝑔 2 (𝑦 𝑜 2 1 , 𝑦 𝑜 2 , … . 𝑦 𝑜 𝑙 ) 𝑦 𝑜+1 = 𝑔 2 (𝑦 𝑜 ….. ….. 𝑗 , 𝑦 𝑜 𝑗 𝑗−1 , 𝑦 𝑜 𝑗+1 ) 𝑦 𝑜+1 = 𝑔 𝑗 (𝑦 𝑜 𝑙 1 , 𝑦 𝑜 2 , … . 𝑦 𝑜 𝑙 ) 𝑦 𝑜+1 = 𝑔 𝑙 (𝑦 𝑜 ….. 𝑙 𝑙−1 , 𝑦 𝑜 𝑙 ) 𝑦 𝑜+1 = 𝑔 𝑙 (𝑦 𝑜 Toroidal boundary conditions ….. 1 𝑙 , 𝑦 𝑜 1 , 𝑦 𝑜 2 ) 1 𝑙 , 𝑦 𝑜 1 , 𝑦 𝑜 2 ) 𝑦 𝑜+1 = 𝑔 1 (𝑦 𝑜 𝑦 𝑜+1 = 𝑔(𝑦 𝑜 ….. ….. Single 𝑗 , 𝑦 𝑜 𝑗 , 𝑦 𝑜 𝑗 𝑗−1 , 𝑦 𝑜 𝑗+1 ) 𝑗 𝑗−1 , 𝑦 𝑜 𝑗+1 ) 𝑦 𝑜+1 = 𝑔 𝑗 (𝑦 𝑜 𝑦 𝑜+1 = 𝑔(𝑦 𝑜 map ….. ….. 𝑙 𝑙−1 , 𝑦 𝑜 𝑙 , 𝑦 𝑜 1 ) 𝑙 𝑙−1 , 𝑦 𝑜 𝑙 , 𝑦 𝑜 1 ) 𝑦 𝑜+1 = 𝑔 𝑙 (𝑦 𝑜 𝑦 𝑜+1 = 𝑔(𝑦 𝑜 4

  5. CA: A F ORMAL DEFINITION  We can give a definition of CAs aside the general framework of DTDS  CAs are defined by:  Components/Cells (Connected FSMs)  Lattice (Geometry + Topology)  Schedule (Time + Synchronization) 5

  6. CA: C OMPONENTS  A set of 𝑁 automata (cells) 𝑏 𝑗 , 𝑗 = 1, … 𝑁 : finite-state machines (in a more general sense, each cell could a function )  Each machine has a specified set of possible states , 𝑇 𝑗 = {𝑡 1 , 𝑡 2 , … , 𝑡 𝑤 }  For each machine 𝑏 𝑗 , state transitions are defined by a local state transition function , that depends on the current state of 𝑏 𝑗 and the state of the 𝑜 𝑗 cells that are in 𝑏 𝑗 ’s neighborhood, 𝒪(𝑏 𝑗 ), 𝐺 𝑗 : 𝑇 𝑗 ∪ 𝒪(𝑏 𝑗 ) → 𝑇 𝑗  At discrete time 𝑢 = 0 , each cell has an initial state , where the vector of all initial states define the initial condition of the CA 𝐺 𝑗 𝑡 𝑗 = 3 𝑏 𝑗 6

  7. CA: L ATTICE  Cells are defined on a lattice , that induces a topology structure  Associated to the topology, is the neighborhood map, 𝒪 , of a cell 𝑏 𝑗 , that associates to 𝑏 𝑗 a set of neighbors , 𝒪 𝑏 𝑗 = {𝑏 𝑘 ∶ 𝑏 𝑘 𝑗𝑡 𝑜𝑓𝑗𝑕ℎ𝑐𝑝𝑠 𝑝𝑔𝑏 𝑗 }  Neighborhood  Range for a cell to be influenced by other cells, range of influence of a cell  Boundary conditions define how the notion of topological neighborhood includes the boundaries, if any, of the lattice  Infinite vs. Finite lattices ( Hard boundaries vs. soft boundaries ) 𝑏 1 𝑏 2 𝑏 3 𝑏 𝑗−1 𝑏 𝑗 𝑏 𝑗+1 1D 𝒪 𝑏 𝑗 = {𝑏 𝑗−1, 𝑏 𝑗+1 } 2D Regular grid Extended Moore Von Neumann Moore 7

  8. CA: L ATTICE , B OUNDARIES  Infinite/adaptive lattice  The grid grows as the pattern propagates  Finite lattice  Hard boundary : fixed, edge cells have a fixed state  Hard boundary : reflective, leftmost (rightmost) cell only diffuse right (left)  Soft boundary: periodic boundary conditions, edges wrap around 8

  9. CA: L ATTICE , B OUNDARIES  Edge wraps around  1D is a ring  2D is torus  Weird(er) topologies with a twist: Moebius bands, Klein bottles 9

  10. CA: S CHEDULES  Synchronized updating: at time 𝑢 the state value of the cells is frozen , and all cells update their state based on their own state and that of their neighbors, then time steps up to 𝑢 + 1 and process is repeated  States are updated in sequence or in parallel, depending on the available hardware, but it doesn’t matter for the final result 𝑢 𝑏 𝑗−2 𝑏 𝑗−1 𝑏 𝑗 𝑏 𝑗+1 𝑏 𝑗+2 𝑢 + 1 𝑏 𝑗−2 𝑏 𝑗−1 𝑏 𝑗 𝑏 𝑗+1 𝑏 𝑗+2 𝐺(𝑏 𝑗 , 𝑏 𝑗−1 , 𝑏 𝑗+1 )  Asynchronous updating: at time 𝑢 the state of one of more cells is updated based on their own state and that of their neighbors at 𝑢 , at 𝑢 + 1 the state of possibly different cells is updated and process is repeated  States are selected according to some criterion, or self-trigger the update, the updating sequence matters for the final result 10

  11. D ESIGN CHOICES  In principle a great freedom choosing:  number and type of states,  state transition functions (for each cell),  topology and neighborhood mapping (for each cell),  cells updating scheme,  number of cells,  boundary conditions  …  In an homogeneous CA, neighborhoods, state transition functions, topology, are the same for all cells, in a non homogenous CA there’s some heterogeneity, in space and/or time, in terms of transitions, topology / neighborhood  Freedom in the design space has been exploited in a number of interesting applications, that precisely might require a diversity of local behaviors , problem- specific interconnection topologies that reflect complex realities such as ecosystems, immune systems, car traffic flows, bio-chemical reactions ,…  CAs are discrete time and space models of partial differential equations 11

  12. D ESIGN CHOICES TO STUDY CA S  Being multidimensional iterated maps, CAs are very complex entities, therefore, to study them, let’s make a few reasonably simplifying assumptions:  Homogeneous CAs:  Lattice is a regular grid, in 1D or 2D  All cell functions 𝑏 have the same (relatively simple) neighborhood mapping 𝒪(𝑏)  they all have the same number of neighbors defined according to the lattice  All cell functions have the same state transition function, 𝐺(𝑏, 𝒪 𝑏 )  States are encoded in a few bits , typically, 2 or 3  Synchronous updating 12

  13. 1D CA  Simplest case: State variables / Cells are Boolean units , 𝑇 = {0,1}  The neighborhood of a cell 𝑏 𝑗 , 𝒪 𝑏 𝑗 corresponds to the one or two closest neighbors in both left and right directions   Transition function 𝐺 is a Boolean function of 𝑜 = 3 or 𝑜 = 5 arguments 𝐺(𝑏 𝑗 , 𝑏 𝑗−1 , 𝑏 𝑗+1 , 𝑏 𝑗−2 , 𝑏 𝑗+2 ) = ቊ1 if 𝑏 𝑗 + 𝑏 𝑗−1 + 𝑏 𝑗+1 + 𝑏 𝑗−2 + 𝑏 𝑗+2 > 2 0 otherwise   A 1D Boolean CA with 𝑜 cells is an 𝑜 -dimensional binary vector 𝒃(t) , the state vector of the CA, that evolves over time by the iterated application of the map 𝐺 : 𝒃 t + 1 = 𝐺(𝒃 t )  State space of the CA : All possible configurations of the vector 𝒃 13

  14. 1D B OOLEAN CA, SOME NUMBERS  𝑙 = 𝑇 = number of (cell) states  𝑇 = {0,1}  𝑙 = 2  𝑁 = number of cells  2 𝑁 possible configurations of CA’s state vector ,  𝑁 = 100, 𝑙 = 2  2 100 ≈ 10 30 !!!!  𝑠 = range = 𝒪 𝑏 /2 (assuming a symmetric neighborhood) One specific function 𝐺  𝑙 2𝑠+1 = 𝑙 |𝒪|+1 possible configurations of neighbor set  If 𝑠 = 1, 𝑙 = 2  8 possible neighbor configurations  If 𝑠 = 2, 𝑙 = 2  32 possible neighbor configurations  𝑙 𝑙 2𝑠+1 = 𝑙 𝑙 |𝒪|+1 = possible evolution functions for the CA 𝒪 = 8  If 𝑠 = 1, 𝑙 = 2  256 possible Boolean evolution functions  If 𝑠 = 2, 𝑙 = 2  4 ∙ 10 9 possible Boolean functions! 14

  15. E LEMENTARY CA: W OLFRAM CODE  𝑇 = 0,1 , 𝑠 = 1  𝑙 = 2, 𝒪 + 1 = 8, 256 possible Boolean functions Transition function 𝑮 (rule of the CA) Example: Rule 30 This is a bit string  Decimal number Rule 30 : (00011110)  30 Wolfram code 15

  16. S OME RULES … 16

  17. S TUDYING CA S : NON - LINEAR BUSINESS AS USUAL Direct problem (Prediction): Given the function, what’s the behavior? 17

  18. R ULE 30 Class 3 cellular automata: overall the evolution presents regularities , however, the state sequence generated by the central cell is used as random generator in Mathematica ! (randomness deriving from a purely deterministic process with no external ’noisy’ inputs) 18

  19. A ZOO OF BEHAVIORS 19

Recommend


More recommend