A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 1 Lower Bounds Lower bound: an estimate on a minimum amount of work needed to solve a given problem Examples: b number of comparisons needed to find the largest element in a set of n numbers b number of comparisons needed to sort an array of size n b number of comparisons necessary for searching in a sorted array b number of multiplications needed to multiply two n-by-n matrices

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 2 Lower Bounds (cont.) b Lower bound can be an exact countan exact count an efficiency class (  )an efficiency class (  ) b Tight lower bound: there exists an algorithm with the same efficiency as the lower bound Problem Lower boundTightness Problem Lower boundTightness sorting  (nlog n) yes sorting  (nlog n) yes searching in a sorted array  (log n) yes searching in a sorted array  (log n) yes element uniqueness  (nlog n) yes element uniqueness  (nlog n) yes n-digit integer multiplication  (n) unknown n-digit integer multiplication  (n) unknown multiplication of n-by-n matrices  (n 2 ) unknown multiplication of n-by-n matrices  (n 2 ) unknown

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 3 Methods for Establishing Lower Bounds b trivial lower bounds b information-theoretic arguments (decision trees) b adversary arguments (SKIP) b problem reduction

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 4 Trivial Lower Bounds Trivial lower bounds: based on counting the number of items that must be processed in input and generated as output Examples b finding max element b polynomial evaluation b sorting b element uniqueness b Hamiltonian circuit existence Conclusions b may and may not be useful b be careful in deciding how many elements must be processed (min-max)

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 5 Decision Trees Decision tree —model algorithms that use comparisons: b internal nodes represent comparisons b leaves represent outcomes b How many orderings of 2 elements? Of 3? Of n? b Tree for 3-element insertion sort [insert b in a, then c into ab]

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 6 Decision Trees and Sorting Algorithms b Any comparison-based sorting algorithm can be represented by a decision tree b Number of leaves (outcomes)  n! b Height of binary tree with n! leaves   log 2 n!  b Minimum number of comparisons in the worst case   log 2 n!  for any comparison-based sorting algorithm b  log 2 3!  =  log 2 6  = 3 b  log 2 n!   n log 2 n, for large n b This lower bound is tight (mergesort)

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 7 Adversary Arguments (SKIP) Adversary argument: a method of proving a lower bound by playing role of adversary that makes algorithm work the hardest by adjusting input Example 1: “Guessing” a number between 1 and n with yes/no questions questions Adversary: Puts the number in a larger of the two subsets generated by last question Example 2: Merging two sorted lists of size n a 1 < a 2 < … < a n and b 1 < b 2 < … < b n a 1 < a 2 < … < a n and b 1 < b 2 < … < b n Adversary: a i < b j iff i < j Output b 1 < a 1 < b 2 < a 2 < … < b n < a n requires 2n-1 comparisons of adjacent elements

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 8 Lower Bounds by Problem Reduction If we know a LB for alg A, can we use that knowledge to find a LB for alg B? Remember reduction: - Minheap reduces to maxheap - Minheap reduces to maxheap - Minimum reduces to sort - Minimum reduces to sort - Element uniqueness for pairs reduces to closest pair - Element uniqueness for pairs reduces to closest pair Which would be most useful? A [  (f )] reduces to B [  (? )] A [  (f )] reduces to B [  (? )] B [  (? )] reduces to A [  (f )] B [  (? )] reduces to A [  (f )] [Assume the reduction itself is fast (ie O(f)) ]

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 9 Lower Bounds by Problem Reduction Which would be most useful? A [  (f )] reduces to B [  (? )] A [  (f )] reduces to B [  (? )] B [  (? )] reduces to A [  (f )] B [  (? )] reduces to A [  (f )] Proc A [  (f )] Proc B [  (? )] … … … … B [  (? )] A [  (f )] B [  (? )] A [  (f )] … … … … A= [  (? )] => ??? A = [  (? )] => ???

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 10 Lower Bounds by Problem Reduction Which would be most useful? A [  (f )] reduces to B [  (? )] A [  (f )] reduces to B [  (? )] B [  (? )] reduces to A [  (f )] B [  (? )] reduces to A [  (f )] Proc A [  (f )] Proc B [  (? )] … … … … B [  (? )] A [  (f )] B [  (? )] A [  (f )] … … … … A= [  (f )] => B = [  (f )] A = [  (f )] => Nothing

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 11 Lower Bounds by Problem Reduction To show lower bounds: Reduce problem with known lower bound to problem with unknown lower bound. Idea: If problem P is at least as hard as problem Q, then a lower bound for Q is also a lower bound for P. bound for Q is also a lower bound for P. Hence, find problem Q with a known lower bound that can be reduced to problem P in question. Hence, find problem Q with a known lower bound that can be reduced to problem P in question. Example: P is finding MST for n points in Cartesian plane Q is element uniqueness problem (known to be in  (nlogn))

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 12 Lower Bounds by Problem Reduction Example: P is finding MST for n points in Cartesian plane Q is element uniqueness problem (known to be in  (nlogn)) Reduce Element Uniqueness to MST: Unique (S : IntSet = {x1, x2, …, xn}) [Known  (nlogn)] Unique (S : IntSet = {x1, x2, …, xn}) [Known  (nlogn)] P : PairSet = {(x1, 0), (x2, 0), …, (xn, 0)} P : PairSet = {(x1, 0), (x2, 0), …, (xn, 0)} T : Tree = MST (P) T : Tree = MST (P) D : Nat = min_length (edges of T) D : Nat = min_length (edges of T) return D = 0 return D = 0 Thus, MST is  (nlogn)! Thus, MST is  (nlogn)!

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 13 P, NP, NP Complete Example: - Classifying Problems - Classifying Problems - Tractable and non-tractable - Optimization and Decision Problems - Problem Classes - Problem Classes - P - NP - NP - NP Complete - NP Complete

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 14 Classifying Problem Complexity Is the problem tractable, i.e., is there a polynomial-time (O(p(n)) algorithm that solves it? Possible answers: b yes (give examples) b no because it’s been proved that no algorithm exists at all (e.g., Turing’s halting problem)because it’s been proved that no algorithm exists at all (e.g., Turing’s halting problem) because it’s been be proved that any algorithm takes exponential timebecause it’s been be proved that any algorithm takes exponential time b unknown

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 15 Problem Types: Optimization and Decision b Optimization problem: find a solution that maximizes or minimizes some objective function b Decision problem: answer yes/no to a question Many problems have decision and optimization versions. E.g.: traveling salesman problem b optimization: find Hamiltonian cycle of minimum length b decision: find Hamiltonian cycle of length  m Decision problems are more convenient for formal investigation of their complexity.

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 16 Class P P: the class of decision problems that are solvable in O(p(n)) time, where p(n) is a polynomial of problem’s input size n Examples: b searching b element uniqueness b graph connectivity b graph acyclicity b primality testing (finally proved in 2002)

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 17 Class NP NP (nondeterministic polynomial): class of decision problems whose proposed solutions can be verified in polynomial time = solvable by a nondeterministic polynomial algorithm A nondeterministic polynomial algorithm is an abstract two-stage procedure that: b generates a random string purported to solve the problem b checks whether this solution is correct in polynomial time By definition, it solves the problem if it’s capable of generating and verifying a solution on one of its tries Why this definition? b led to development of the rich theory called “computational complexity”

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 18 Example: CNF satisfiability Problem: Is a boolean expression in its conjunctive normal form (CNF) satisfiable, i.e., are there values of its variables that makes it true? This problem is in NP. Nondeterministic algorithm: b Guess truth assignment b Substitute the values into the CNF formula to see if it evaluates to true Example: ( A | ¬B | ¬C ) & ( A | B ) & ( ¬B | ¬D | E ) & ( ¬D | ¬E ) Truth assignments: A B C D E Checking phase: O(n)

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 19 What problems are in NP? b Hamiltonian circuit existence b Partition problem: Is it possible to partition a set of n integers into two disjoint subsets with the same sum? b Decision versions of TSP, knapsack problem, graph coloring, and many other combinatorial optimization problems. (Few exceptions include: MST, shortest paths) b All the problems in P can also be solved in this manner (no guessing is necessary), so we have: P  NP P  NP  Big question: Is P = NP ? Or is P ⊂ NP ? Answer unknown. Current approach: focus on hardest NP problemsAnswer unknown. Current approach: focus on hardest NP problems

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 20 NP-Complete : The Hardest NP Problems A decision problem D in NP is NP-complete if it’s as hard as (ie no easier than) any and every problem in NP. That is: b D is in NP b every problem in NP is polynomial-time reducible to D Think: Which problems can be easier?

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 21 Cook: The First NP-Complete Problem A decision problem D that is in NP is NP-complete if b D is in NP b Every problem in NP is polynomial-time reducible to D How to prove every NP problem is reducible to D? Not easy! Cook’s theorem (1971): CNF-sat is NP-complete

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 22 Proving Other Problems are NP-Complete Prove other problems are NP-complete by polynomial- time reductions from a known NP-complete problem Pay attention to the direction of the reductions Examples: TSP, knapsack, partition, graph-coloring and hundreds of other problems of combinatorial nature

A. Levitin “Introduction to the Design & Analysis of Algorithms,” 3rd ed., Ch. 11 ©2012 Pearson Education, Inc. Upper Saddle River, NJ. All Rights Reserved. 23 P = NP ? Dilemma Revisited b If P = NP then every problem in NP (including all NP- complete problems) could be solved in polynomial time b If a polynomial-time algorithm for just one NP-complete problem is discovered, then every problem in NP can be solved in polynomial time, i.e., P = NP b Most but not all researchers believe that P  NP, i.e. P is a proper subset of NP