Permutations 2 Marion Scheepers Permutations 2 Marion Scheepers Boise State University Summer 2012 Permutations review Permutations 2 Marion Scheepers Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : The set of permutations of {1, 2, · · · , n}. Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : The set of permutations of {1, 2, · · · , n}. Fundamental Fact 1: Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : The set of permutations of {1, 2, · · · , n}. Fundamental Fact 1:Under functional composition, Sn is a group. Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : The set of permutations of {1, 2, · · · , n}. Fundamental Fact 1:Under functional composition, Sn is a group. Fundamental Fact 2 (Cayley’s Theorem): Permutations review Permutations 2 Marion Scheepers f = 1 f(1) 2 f(2) ··· ··· n-1 f(n-1) n f(n) Permutation: A one-to-one function from a finite set S to S. Sn : The set of permutations of {1, 2, · · · , n}. Fundamental Fact 1:Under functional composition, Sn is a group. Fundamental Fact 2 (Cayley’s Theorem): Every finite group is a subgroup of some Sn . Reversals Permutations 2 Marion Scheepers Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps 1 2 ··· i-1 i i+1 ··· j-1 j j+1 ··· n Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps 1 2 ··· i-1 i i+1 ··· to j-1 j j+1 ··· n Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps 1 2 ··· i-1 i i+1 ··· j-1 j j+1 ··· n i+1 i j+1 ··· n to 1 2 ··· i-1 j j-1 ··· Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps 1 2 ··· i-1 i i+1 ··· j-1 j j+1 ··· n i+1 i j+1 ··· n to 1 2 ··· i-1 j is said to be a reversal. j-1 ··· Reversals Permutations 2 Marion Scheepers For fixed i < j in {1, 2, · · · , n} the function f that maps 1 2 ··· i-1 i i+1 ··· j-1 j j+1 ··· n i+1 i j+1 ··· n to 1 2 ··· i-1 j j-1 ··· is said to be a reversal. With i = 1, it is said to be a prefix reversal. Reversals 2 Permutations 2 Marion Scheepers Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . The reversal distance problem Given positive integer n, Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . The reversal distance problem Given positive integer n, and given permutations f and g from Sn , Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . The reversal distance problem Given positive integer n, and given permutations f and g from Sn , what is the length of the shortest sequence ρ1 , · · · , ρk of reversals such that Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . The reversal distance problem Given positive integer n, and given permutations f and g from Sn , what is the length of the shortest sequence ρ1 , · · · , ρk of reversals such that f = ρk ◦ · · · ◦ ρ1 ◦ g ? Reversals 2 Permutations 2 Marion Scheepers The subset Rn of Sn consisting of the reversals generates Sn . The reversal distance problem Given positive integer n, and given permutations f and g from Sn , what is the length of the shortest sequence ρ1 , · · · , ρk of reversals such that f = ρk ◦ · · · ◦ ρ1 ◦ g ? Theorem: (Even and Goldreich, 1981) The reversal distance problem is NP-complete. Reversals 3: Gollan’s Conjecture Permutations 2 Marion Scheepers Reversals 3: Gollan’s Conjecture Permutations 2 Marion Scheepers Conjecture: (Gollan) The maximum reversal distance for Sn is n − 1. Reversals 3: Gollan’s Conjecture Permutations 2 Marion Scheepers Conjecture: (Gollan) The maximum reversal distance for Sn is n − 1. Theorem: (Bafna and Pevzner, 1996) Gollan’s conjecture is true. Constrained Reversals Permutations 2 Marion Scheepers Constrained Reversals Permutations 2 Marion Scheepers Fundamental Fact: The set Pn of prefix reversals of Sn generates Sn . Constrained Reversals Permutations 2 Marion Scheepers Fundamental Fact: The set Pn of prefix reversals of Sn generates Sn . Theorem (Gates and Papadimitriou, 1978) For each positive integer n, it requires at most 5n+5 prefix reversals to generate 3 an element of Sn Constrained Reversals Permutations 2 Marion Scheepers Fundamental Fact: The set Pn of prefix reversals of Sn generates Sn . Theorem (Gates and Papadimitriou, 1978) For each positive integer n, it requires at most 5n+5 prefix reversals to generate 3 an element of Sn Theorem (Even and Goldreich, 1981) The prefix reversal distance problem is NP-complete. Constrained Reversals Permutations 2 Marion Scheepers Fundamental Fact: The set Pn of prefix reversals of Sn generates Sn . Theorem (Gates and Papadimitriou, 1978) For each positive integer n, it requires at most 5n+5 prefix reversals to generate 3 an element of Sn Theorem (Even and Goldreich, 1981) The prefix reversal distance problem is NP-complete. Theorem (Gates and Papadimitriou, 1978) For each positive integer multiple n of 16, there is a member of Sn that is not generated by fewer than 17n 16 prefix reversals. Block Swaps Permutations 2 Marion Scheepers Block Swaps Permutations 2 Marion Scheepers For fixed i ≤ j < k ≤ m in {1, 2, · · · , n} Block Swaps Permutations 2 Marion Scheepers For fixed i ≤ j < k ≤ m in {1, 2, · · · , n} the function f that swaps Block Swaps Permutations 2 Marion Scheepers For fixed i ≤ j < k ≤ m in {1, 2, · · · , n} the function f that swaps the blocks i i + 1 ···j and k k + 1 ···m Block Swaps Permutations 2 Marion Scheepers For fixed i ≤ j < k ≤ m in {1, 2, · · · , n} the function f that swaps the blocks i i + 1 ···j and k k + 1 ···m and leaves the rest of its domain unchanged, is a block interchange. Block Swaps Permutations 2 Marion Scheepers For fixed i ≤ j < k ≤ m in {1, 2, · · · , n} the function f that swaps the blocks i i + 1 ···j and k k + 1 ···m and leaves the rest of its domain unchanged, is a block interchange. Fundamental Fact: The set Bn of block interchanges of Sn generates Sn . Block Swaps 2 Permutations 2 Marion Scheepers Block Swaps 2 Permutations 2 Marion Scheepers Theorem: (Christie, 1996) For a positive integer n: Block Swaps 2 Permutations 2 Marion Scheepers Theorem: (Christie, 1996) For a positive integer n: (a) There are elements of Sn which cannot be represented by fewer than b n2 c block interchange operations. Block Swaps 2 Permutations 2 Marion Scheepers Theorem: (Christie, 1996) For a positive integer n: (a) There are elements of Sn which cannot be represented by fewer than b n2 c block interchange operations. (b) Every element of Sn is representable by at most b n2 c block interchanges. Block Swaps 2 Permutations 2 Marion Scheepers Theorem: (Christie, 1996) For a positive integer n: (a) There are elements of Sn which cannot be represented by fewer than b n2 c block interchange operations. (b) Every element of Sn is representable by at most b n2 c block interchanges. (c) A sequence of block interchanges of minimal length representing an element of Sn can be found in time polynomial in n. Block Swaps 2 Permutations 2 Marion Scheepers Theorem: (Christie, 1996) For a positive integer n: (a) There are elements of Sn which cannot be represented by fewer than b n2 c block interchange operations. (b) Every element of Sn is representable by at most b n2 c block interchanges. (c) A sequence of block interchanges of minimal length representing an element of Sn can be found in time polynomial in n. Thus, the Block swap distance problem is solvable in polynomial time. An illustration. (n=11) Permutations 2 Marion Scheepers An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 4 6 8 7 9 2 1 10 11 An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 4 6 8 7 9 2 1 10 11 Step 2: 3 4 5 6 8 7 9 2 1 10 11 An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 4 6 8 7 9 2 1 10 11 Step 2: 3 4 5 6 8 7 9 2 1 10 11 Step 3: 3 4 5 6 7 8 9 2 1 10 11 An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 4 6 8 7 9 2 1 10 11 Step 2: 3 4 5 6 8 7 9 2 1 10 11 Step 3: 3 4 5 6 7 8 9 2 1 10 11 Step 4: 9 8 7 6 5 4 3 2 1 10 11 An illustration. (n=11) Permutations 2 Rearranging 3 5 8 6 4 7 9 2 1 10 11 using reversals: Marion Scheepers Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 4 6 8 7 9 2 1 10 11 Step 2: 3 4 5 6 8 7 9 2 1 10 11 Step 3: 3 4 5 6 7 8 9 2 1 10 11 Step 4: 9 8 7 6 5 4 3 2 1 10 11 Step 5: 1 2 3 4 5 6 7 8 9 10 11 An illustration. (n=11) Permutations 2 Marion Scheepers An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using block swaps: An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using block swaps: Step 0: 3 5 8 6 4 7 9 2 1 10 11 An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using block swaps: Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 6 4 7 8 9 2 1 10 11 An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using block swaps: Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 6 4 7 8 9 2 1 10 11 Step 2: 1 2 3 5 6 4 7 8 9 10 11 An illustration. (n=11) Permutations 2 Marion Scheepers Rearranging 3 5 8 6 4 7 9 2 1 10 11 using block swaps: Step 0: 3 5 8 6 4 7 9 2 1 10 11 Step 1: 3 5 6 4 7 8 9 2 1 10 11 Step 2: 1 2 3 5 6 4 7 8 9 10 11 Step 3: 1 2 3 4 5 6 7 8 9 10 11 The break point graph Permutations 2 Marion Scheepers The break point graph Permutations 2 Marion Scheepers Given: The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. a0 = 0, an+1 = n + 1. The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. a0 = 0, an+1 = n + 1. Eλb = {{ai , aj } : (ai , aj ) or (aj , ai ) a breakpoint or adjacent in λ} The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. a0 = 0, an+1 = n + 1. Eλb = {{ai , aj } : (ai , aj ) or (aj , ai ) a breakpoint or adjacent in λ} Eλg = {{ai , aj } : |i − j| > 1 but |aj − ai | = 1} The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. a0 = 0, an+1 = n + 1. Eλb = {{ai , aj } : (ai , aj ) or (aj , ai ) a breakpoint or adjacent in λ} Eλg = {{ai , aj } : |i − j| > 1 but |aj − ai | = 1} S Eλ = Eλb Eλg , the edge set. The break point graph Permutations 2 Marion Scheepers Given: a repetition free listing λ = a1 a2 · · · an of {1, · · · , n}. For 1 ≤ i < n, (ai , ai+1 ) is a break point if |ai − ai+1 | > 1. b(λ) : The number of break points of λ Define a graph: V (λ) = {0, 1, · · · , n, n + 1}, the vertex set. a0 = 0, an+1 = n + 1. Eλb = {{ai , aj } : (ai , aj ) or (aj , ai ) a breakpoint or adjacent in λ} Eλg = {{ai , aj } : |i − j| > 1 but |aj − ai | = 1} S Eλ = Eλb Eλg , the edge set. (V , Eλ ) is the break point graph of λ. The break point graph 2 Permutations 2 Marion Scheepers The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: Each {xi , xi+1 } and {x1 , xn } are edges in the break point graph and The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: Each {xi , xi+1 } and {x1 , xn } are edges in the break point graph and the path goes through alternating “b” and “g” edges. The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: Each {xi , xi+1 } and {x1 , xn } are edges in the break point graph and the path goes through alternating “b” and “g” edges. c(λ) The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: Each {xi , xi+1 } and {x1 , xn } are edges in the break point graph and the path goes through alternating “b” and “g” edges. c(λ) The maximum number of edge-disjoint cycles that cover the vertex set of the break point graph. The break point graph 2 Permutations 2 Marion Scheepers A cycle in the breakpoint graph: (x1 , x2 , · · · , xn ) where: Each {xi , xi+1 } and {x1 , xn } are edges in the break point graph and the path goes through alternating “b” and “g” edges. c(λ) The maximum number of edge-disjoint cycles that cover the vertex set of the break point graph. The break point graph 3 Permutations 2 Marion Scheepers The break point graph 3 Permutations 2 Marion Scheepers Theorem (Bafna and Pevzner, 1996) The reversal distance for a permutation f ∈ Sn is at least n + 1 − c(f ). The cycle graph Permutations 2 Marion Scheepers The cycle graph for [f1 f2 · · · fn−1 fn ] is a directed graph defined to have vertex set {0, 1, · · · , n, n + 1} where we set f0 = 0 and fn+1 = n + 1, and directed edge set Efb = {(fi+1 , fi ) : 0 ≤ i ≤ n} and Efg = {(i, i + 1) : 0 ≤ i ≤ n}. The cycle graph Permutations 2 Marion Scheepers The cycle graph for [f1 f2 · · · fn−1 fn ] is a directed graph defined to have vertex set {0, 1, · · · , n, n + 1} where we set f0 = 0 and fn+1 = n + 1, and directed edge set Efb = {(fi+1 , fi ) : 0 ≤ i ≤ n} and Efg = {(i, i + 1) : 0 ≤ i ≤ n}. For permutation f , cg (f ) denotes the number of alternating cycles in the cycle graph of f . Theorem (Christi, 1996) The block interchange distance for a permutation f ∈ Sn is 21 (n + 1 − cg (f )). Exercises Permutations 2 Marion Scheepers Exercises Permutations 2 Marion Scheepers 1. Construct the break point graphs for the examples given in these talks. Exercises Permutations 2 Marion Scheepers 1. Construct the break point graphs for the examples given in these talks. 2. Compute the quantities b(λ) and c(λ) for these examples. Exercises Permutations 2 Marion Scheepers 1. Construct the break point graphs for the examples given in these talks. 2. Compute the quantities b(λ) and c(λ) for these examples. 3. Construct the cycle graphs for the examples given in these talks. Exercises Permutations 2 Marion Scheepers 1. Construct the break point graphs for the examples given in these talks. 2. Compute the quantities b(λ) and c(λ) for these examples. 3. Construct the cycle graphs for the examples given in these talks. 4. Compute the quantity cg (λ) for these examples. Exercises Permutations 2 Marion Scheepers 1. Construct the break point graphs for the examples given in these talks. 2. Compute the quantities b(λ) and c(λ) for these examples. 3. Construct the cycle graphs for the examples given in these talks. 4. Compute the quantity cg (λ) for these examples. Upcoming talk Permutations 2 Marion Scheepers Upcoming talk Permutations 2 Marion Scheepers Upcoming talk Permutations 2 Marion Scheepers 1. Combining reversals and block swaps. Upcoming talk Permutations 2 Marion Scheepers 1. Combining reversals and block swaps. 2. Another set of constrained reversals and constrained block swaps. Upcoming talk Permutations 2 Marion Scheepers 1. Combining reversals and block swaps. 2. Another set of constrained reversals and constrained block swaps. 3. Open problems. Literature Permutations 2 Marion Scheepers Literature Permutations 2 Marion Scheepers V. Bafna and P.A. Pevzner, Genome rearrangements and sorting by reversals, SIAM Journal of Computation 25:2 (1996), 272 - 289. Literature Permutations 2 Marion Scheepers V. Bafna and P.A. Pevzner, Genome rearrangements and sorting by reversals, SIAM Journal of Computation 25:2 (1996), 272 - 289. D.A. Christie, Sorting permutations by block interchanges, Information Processing Letters 60 (1996), 165 - 169. Literature Permutations 2 Marion Scheepers V. Bafna and P.A. Pevzner, Genome rearrangements and sorting by reversals, SIAM Journal of Computation 25:2 (1996), 272 - 289. D.A. Christie, Sorting permutations by block interchanges, Information Processing Letters 60 (1996), 165 - 169. W.H. Gates and C.H. Papadimitriou, Bounds for sorting by prefix reversal, Discrete Mathematics 27 (1979), 47 - 57. Literature Permutations 2 Marion Scheepers V. Bafna and P.A. Pevzner, Genome rearrangements and sorting by reversals, SIAM Journal of Computation 25:2 (1996), 272 - 289. D.A. Christie, Sorting permutations by block interchanges, Information Processing Letters 60 (1996), 165 - 169. W.H. Gates and C.H. Papadimitriou, Bounds for sorting by prefix reversal, Discrete Mathematics 27 (1979), 47 - 57. S. Even and O. Goldreich, The minimum length generator sequence problem is NP-hard, Journal of Algorithms 2 (1981), 311 - 313.