the maximum clique problem mcp
play

The Maximum Clique Problem (MCP) You are given: An undirected - PowerPoint PPT Presentation

The Maximum Clique Problem (MCP) You are given: An undirected graph G = ( V , E ) , where - V = {1,., n } - E V x V and are asked to Find the largest complete subgraph (clique) of G The problem is known to be NP-hard, and so is


  1. The Maximum Clique Problem (MCP) You are given: • An undirected graph G = ( V , E ) , where - V = {1,…., n } ⊆ - E V x V and are asked to • Find the largest complete subgraph (clique) of G The problem is known to be NP-hard, and so is problem of determining just the size of the maximum clique. Pardalos and Xue (1994) provide a review of the MCP with 260 references.

  2. The Maximum Clique Problem (MCP) Affrontando il problema MCP in termini di rete neurale: • Trasformare MCP da problema discreto a problema continuo Nell’esempio del TSP con il modello di Hopfield, non è detto che ci sia il percorso inverso (potremmo ottenere ad esempio una matrice che non ha significato); in questo nuovo problema MCP, la bidirezionalità è d’obbligo.

  3. Some Notation Given an arbitrary graph G = ( V , E ) with n nodes: If C V , x c will denote its characteristic vector which is defined as ⊆ •  ∈ 1 C , if i C = x c  i  0 , otherwise S n is the standard simplex in R n : •   n ∑ = ∈ = ≥ ∀ n   S x R x and x i : 1 0 , n i i   = i 1 • A =(a ij ) is the adjacency matrix of G :  1 , if v ~ v = i j  a ij  otherwise 0 ,

  4. Infeasible Maxima in Motzkin-Straus Si consideri la funzione continua: = ∑ ∑ n n ( ) ′ = f x x A x a x x ij i j = = i 1 j 1 x ' dove è il vettore trasposto e A è la matrice di adiacenza. Lagrangiano del grafo: ( ) = ∑ f x x x i j ∈ i j , E esempio: 1 ( ) = + f x x , , x x x x x 1 2 3 1 2 1 3 2 3

  5. Continuous Formulation of MAX-CLIQUE Il ponte che crea Motzkin-Straus è unidirezionale; solo se il vettore restituito è nella forma di vettore caratteristico allora c’è bidirezionalità. Nell’esempio visto ci sono due massimi globali : T T     1 1 1 1 ′ ′′ = =     x , ,0 x ,0,     2 2 2 2 Si dimostra che sono massimi globali anche tutti i punti del segmento x’ - x’’ α − α T   1 1 [ ] ∀ α ∈   , , 0,1 ovvero tutti i punti ; non essendo vettori caratteristici   2 2 2 (soluzioni spurie) non è possibile estrarre la clique massima. La soluzione con- 1 siste nel sommare alla diagonale principale di A 2   1 ( ) ( ) 1 = + ′ ′ = = T + T   f x x A I x f x x A x A A I   2 2

  6. Infeasible Maxima in Motzkin-Straus Teorema V e x c vettore caratteristico allora: ⊆ Dato C - C è una clique massima di G x c è un massimo globale di in f S n f - C è una clique massimale di G x c è un massimo locale di in S n - tutti i massimi locali sono stretti e sono vettori caratteristici

  7. Evolutionary Games Developed in evolutionary game theory to model the evolution of behavior in animal conflicts. Assumptions • A large population of individuals belonging to the same species which compete for a particular limited resource • This kind of conflict is modeled as a game, the players being pairs of randomly selected population members • Players do not behave “rationally” but act according to a pre-programmed behavioral pattern, or pure strategy • Reproduction is assumed to be asexual • Utility is measured in terms of Darwinian fitness, or reproductive success

  8. Notations { } = • is the set of pure strategies J 1, L , n ( ) t x t • i is the proportion of population members playing strategy at time i x ′ ( ) = • x x 1 , L , The state of population at a given instant is the vector n ( ) σ • x x x Given a population state , the support of , denoted , is defined as x the set of positive components of , i.e., ( ) { } σ = ∈ > x i J x : 0 i

  9. Payoffs ( ) × = n n Let be the payoff (or fitness) matrix. A a ij a i represents the payoff of an individual playing strategy against an opponent ij j ( ) ∈ playing strategy . i j , J x i If the population is in state , the expected payoff earnt by an – strategist is: n ∑ ( ) ( ) π = = x a x Ax i ij j i = j 1 while the mean payoff over the entire population is: n ∑ ( ) ( ) ′ π = π = x x x x Ax i i = i 1

  10. Replicator Equations Developed in evolutionary game theory to model the evolution of behavior in animal conflicts (Hofbauer & Sigmund, 1998; Weibull, 1995). ( ) × = Let be a non-negative real-valued n n matrix, and let W w ij = ∑ n ( ) ( ) π t w x t i ij j = j 1 Continuous-time version:   n d x t ( ) ( ) ( ) ∑ ( ) ( ) = π − π   x t t x t t   i i i j j dt   = j 1 Discrete-time version: ( ) ( ) π x t t ( ) + = i i x t 1 ∑ ( ) ( ) i n π x t t = j j j 1

  11. Replicator Equations & Fundamental Theorem of Selection S is invariant under both dynamics, and they have the same stationary points. n W ′ = Theorem: If , then the function W ( ) ′ = F x x W x is strictly increasing along any non-constant trajectory of both continuous-time and discrete-time replicator dynamics

  12. Mapping MCP’s onto Relaxation Nets n To (approximately) solve a MCP by relaxation, simply construct a net having { } units, and a -weight matrix given by 0,1 1 = + W A I n 2 where A is the adjacency matrix of G. Example:

  13. Mapping MCP’s onto Relaxation Nets The system starting from u(0) will maximize the Motzkin-Straus function and will converge to a fixed point u * which corresponds to a (local) maximum of . f The value 1 ∗ = k ( ) ∗ − 1 2 f u can be regarded as an approximation of the maximum clique size. Con –measure si misura la qualità Q − f f = ave RE Q − α f ave f f dove è il termine di confronto rispetto alla media, è la replicator ave RE α Q equation e è il valore ottimale. Quando 1 il risultato è buono.

  14. Experimental Setup Experiments were conducted over random graphs having: n • size: = 10, 25, 50, 75, 100 δ • density: = 0.10, 0.25, 0.50, 0.75, 0.90 Comparison with Bron-Kerbosch (BK) clique-finding algorithm (1974). δ n δ For each pair ( , ) 100 graphs generated randomly with size and density ≈ n . δ n The case = 100 and = 0.90 was excluded due to the high cost of BK algorithm. Total number of graphs = 2400. n 10 25 50 75 100 δ 0.10 0.99 (54) 0.99 (36) 0.99 (53) 0.97 (59) 0.92 (82) 0.25 0.99 (54) 0.99 (64) 0.99 (84) 1.00 (98) 0.97 (112) 0.50 1.00 (56) 0.99 (118) 0.99 (153) 0.96 (160) 0.90 (187) 0.75 1.00 (99) 1.00 (175) 1.00 (268) 1.00 (284) 1.00 (369) 0.90 1.00 (119) 1.00 (224) 1.00 (367) 0.99 (513) ---- Values of Q-measure for various sizes and densities

Recommend


More recommend