CMU-Q Lecture 2: Search problems Uninformed search. Teacher: Gianni A. Di Caro
|
|
- Osborn Sparks
- 5 years ago
- Views:
Transcription
1 CMU-Q Lecture 2: Search problems Uninformed search Teacher: Gianni A. Di Caro
2 RECAP: ACT RATIONALLY Think like people Think rationally Agent Sensors? Actuators Percepts Actions Environment Act like people Act rationally Techniques for rational decisionmaking in machines Action a Abstraction Problem State A State B Formal Model Problem Class 2
3 RECAP: TASK ENVIRONMENT~PROBLEM Representation (States, Transitions) Agent Sensors? Actuators Percepts Actions Environment Agent behavior: Agent function that maps percept sequences to an action Task environment (~Problem): Environment + Actuators + Sensors + Performance Different task environments require different techniques for problem representation, agent design, and rational decision-making 3
4 4
5 (MODEL-BASED) AUTOMATED PLANNING Task environment: Fully Observable Deterministic Known Sequential Discrete Single agent Objective: The agent has well defined goals that aims to reach at the minimum cost, starting from a defined start state Output: A plan that takes from the start state to a goal state at the minimum cost Plan: A sequence of decisions to be executed from start state in open-loop Problem representation: A graph, where nodes are problem states and arcs are feasible actions (state transitions) 5
6 GOAL-BASED AGENT Has goal(s), that describe situations that are desirable The agent uses goal information to select between possible actions in the current state by accounting for their effect in the future toward the achievement of goals Search: the process of looking for a sequence of actions that reaches the goal (optimizing performance) 6
7 SEARCH / DETERMINISTIC PLANNING A search problem is defined by: ü States (configurations of the environment) ü Start state and goal state(s) ü Actions available to the agent in each state ü Transition model: the state resulting from doing action a in state s (Successor function) ü Cost model: step costs c(s, a, s & ) 0, path costs (additive) What are the states? Find one (the optimal) sequence of actions (path) from start to a goal state 7
8 PATH SEARCH ON GRAPHS State space of the problem: Set of all states reachable from initial state by any sequence of actions Determined by: initial state + actions + transition model Ø The state space structure can be represented as a directed graph where the nodes are states and edges between nodes are actions1 1 Minimum cost path finding on graphs 8
9 EXAMPLE: PANCAKE FLIPPING s t Finding the minimum number of flips is NP-hard 1 1 Number of states? 9
10 EXAMPLE: 8-PUZZLE s t States Locations of tiles 8-puzzle: 181,440 states 15-puzzle: 1.3 trillion states 24-puzzle: states Actions Move blank left, right, up, down NP-Complete! Path cost 1 per move 10
11 EXAMPLE: PATHFINDING Initial state Arad Actions Go from one city to another Transition model If you go from city A to city B, you end up in city B Goal state Bucharest Path cost Sum of edge costs (traveling times, Km) 11
12 EXAMPLE: TOURING PROBLEMS Visit every city at least once, starting and ending in Bucharest How the state space looks like? Traveling salesperson problem (TSP): Visit each city exactly once and find the tour of minimum cost (costs may be not symmetric ) # states for a symmetric n-cities TSP: 12
13 EXAMPLE: TOURING 16,892 cities PROBLEMS 71,009 cities Trivial to find one feasible solution, (NP-) hard to find the best one 13
14 SOLUTION APPROACHES FOR SEARCH Meta-Algorithms: incremental node generation Ø Tree search Ø Graph search Uninformed ü Breadth-First Search (BFS) ü Bidirectional BFS Ø Uniform-Cost Search (UCS) Informed Greedy Best-First Search v A* Ø Theta* Depth-First Search (DFS) Depth-Limited DFS Ø Iterative Deepening Search (IDS) 14
15 TREE SEARCH o Let s begin at the start node and expand it: make the set of all its possible successor states (i.e., resulting from feasible actions) o Include the set resulting from the node expansion into a set of unexpanded nodes, that is called the frontier set Successor state Starting state Action Frontier set o At each step, pick a node from the frontier to expand o Keep going until reach a goal state o Try to expand as few nodes as possible, or move along the path of minimum cost Goal state Search tree 15
16 TREE SEARCH function TREE-SEARCH(problem, strategy) set of frontier nodes containing the start state of problem loop if there are no frontier nodes then return failure else choose a frontier node for expansion using strategy if the node contains a goal state then return the corresponding solution else expand the node and add the resulting nodes to the set of frontier nodes end loop 16
17 TREE SEARCH EXAMPLE 17
18 TREE SEARCH EXAMPLE 18
19 TREE SEARCH EXAMPLE Strategy Frontier 19
20 TREE SEARCH Frontier nodes Explored nodes 20
21 TREE SEARCH ISSUES Tree search can expand the same states again and again Algorithms that forget their history are doomed to repeat it! 21
22 ANOTHER CYCLING EXAMPLE Frontier/open set 1 {m} 2 {h,l,r,n} 3 {l,r,n,c,g,i,m} 4 {l,r,n,c,i,m,b,f,h} 5 {l,r,c,i,m,b,f,h,o,s} 6 {l,r,c,m,b,f,h,o,s,d,n,j} 7 Cycle! c 4 b c g 2 h 1 m 3 g i m f h i 5 l r n 6 i m o 7 d h n j s a f k p u b g l q v c h m r w Goal state a can be reached from start state m in multiple ways. The same intermediate state c can be encountered following different paths from m d i n s x e j o t y 22
23 STATES VS. NODES State s: an admissible configuration of the world Node n: a bookkeeping data structure used to represent the search tree o Nodes are on specific search paths, as defined by PARENT pointers, states are not o Two different nodes on the tree can contain the same state s, if s was generated via two different search paths Right PARENT cost-to come start STATE Node ACTION = Right PATH-COST =
24 STATES VS. NODES A Node s data structure contains: n.state: the state s to which n corresponds to n.parent: the node in the search tree that generated node n n.action: the action that was applied to the parent to generate n n.path-cost: the cost g(n) of the path from the initial node to n as indicated by the parent pointers (cost-to-come) start Right PARENT cost-to come STATE Node ACTION = Right PATH-COST = 6 Same state, different nodes 24
25 GRAPH SEARCH function GRAPH-SEARCH(problem, strategy) set of frontier nodes containing the start state of problem loop if there are no frontier nodes then return failure else choose a frontier node for expansion using strategy, and add it to the explored set if the node contains a goal state then return the corresponding solution else expand the node and add the resulting nodes to the set of frontier nodes, only if not in the frontier or explored set 25
26 GRAPH SEARCH ILLUSTRATED Each node is associated to a different state, the search tree grows on the state graph! Separation property: Every path from initial state to an unexplored state has to pass through the frontier. The frontier separates explored vs. unexplored regions 26
27 NO REPEATED STATES IN GRAPH SEARCH a f k p u b g l q v c h m r w d i n s x e j o t y 1 m 2 5 h l r n c 3 6 g i m i m o s 4 7 b f h i d h n j a c 8 g c e i 9 b 10 d h a c g Path to goal from start: (a b c d i n m) Cost: 6 Additional data structure to hold explored/closed set + a map function to search it (e.g., hash function) 27
28 MEASURING PERFORMANCE How do we will score the performance of different approaches for finding a solution? Completeness Optimality Time Space Guaranteed to find a solution when there is one? Finds the best solution? How long does it take to find a solution? How much memory needed to perform the search? 28
29 MEASURING PERFORMANCE o In AI scenarios, state graphs are usually HUGE, such that they are given implicitly, trough the initial state, actions and the transition model (the configuration resulting from applying action a to state s). o Graph nodes are generated lazily during the search process based on node expansion choices A B E C D Explicit graph description Tree search generation example: A {B,C} B {A,C} C {A,B,D} B {A,C,D} D {A,C,E} 29
30 MEASURING PERFORMANCE It may be not always useful/possible to reason on V and E to quantify performance, but rather other quantities are considered Branching factor b, the maximum number of successors of any node Depth d, the depth (# of steps along path from root) of shallowest goal Max length, of any path from start in the state space # nodes generated during search ~ Time Max # of nodes stored in memory during search ~ Space 30
31 UNINFORMED VS. INFORMED Based on either tree or graph search (or both), two main classes of search algorithms exist, which mainly differentiate from each other because of the strategy used to expand nodes : Uninformed Can only generate successors, sum up what has happened so far, and distinguish goals from non-goals Informed Strategies that know whether one non-goal is more promising than another 31
32 UNINFORMED VS. INFORMED SEARCH Strategy How desirable is to be in a certain intermediate state for the sake of (effectively) reaching a goal state from start Uninformed: Generation order, estimated cost-to-come Informed: Generation order, estimated cost-to-come, estimated cost-to-go Heuristic information Start ~ Cost-to-come (Pessimistic) ~ Cost-to-go (Optimistic) Goal 32
33 UNINFORMED SEARCH Uninformed Can only generate successors, use cost-to-come, and distinguish goals from non-goals 33
34 BREADTH-FIRST SEARCH (BFS) Strategy: Expand shallowest unexpanded node All nodes at a given level of the search tree are expanded before any other node is expanded. Horizontal, level-by-level search Can be implemented by using a FIFO queue for the frontier set Goal test applied when a node is generated 34
35 BFS FOR PATH FINDING 35
36 BFS GRAPH SEARCH FOR 8-PUZZLE (One) Goal found after 26 node expansions and 47 goal checks 36
37 BFS PROPERTIES Algorithm BFS Complete? Optimal? Time Space ~Yes Not really Θ b 5 Θ b 5 Complete: Yes if a goal node is a finite depth. It s true also in infinite graphs. In both cases, the branching factor b must be finite Optimality: Only if costs are a non-decreasing function of the depth (e.g., all actions have the same cost). In general, the shallowest goal (the first generated) is not necessarily the optimal one As it is, not useful for finding Shortest Paths
38 BFS PROPERTIES Algorithm BFS Complete? Optimal? Time Space ~Yes Not really Θ b 5 Θ b 5 Time complexity: Imagine each node has exactly b successors, and solution is at depth d, then BFS will generate 5 8:; b 8 = Θ b 5 nodes. b=2 d=7 Space complexity: For graph search BFS, Θ b 5 nodes are in frontier and Θ b 5<; in explored set ( tree search would not save much, since only the explored nodes could be avoided). E.g., for b=10, d=12 Θ(10 12 ) nodes. If 1 node requires 1 kb kb! 38
39 BIDIRECTIONAL SEARCH Idea: Possibly improve the running time of BFS by running two simultaneous searches, forward from the initial state and backward from the goal What is the worst-case running time of BIDIRECTIONAL SEARCH? 1. Θ(b d) 2. Θ((b/2) 5 ) 3. Θ(b 5/C ) 4. Θ(b 5 ) 39
40 BIDIRECTIONAL SEARCH Θ(b 5/C ) + Θ(b 5/C F ) 2 b 5 Θ(b 5 ) For b=10, d=6, each BFS generates up to depth d=3 2,220 nodes vs. 1,111,110: big memory save! Time complexity: Θ(b 5/C ) running the two searches in parallel, or Θ(b 5 ) if sequential alternating Issues: Asymmetric costs Unidirectional moves Repeated check for frontier intersection (additional constant time with hashing) Existence of multiple goals Abstract goal definition 40
41 BID-BFS: PATH FINDING EXAMPLE 41
42 UNIFORM-COST SEARCH (UCS) Strategy: Expand frontier node with lowest (estimated) path cost g n from the starting node, the cost-to-come Can be implemented by using a priority queue (PQ) ordered by g(n) for the frontier set Other changes from BFS: o o o Goal test applied when node is selected for expansion, not when included in the frontier (the first time a node is selected it can be on a sub-optimal path) If a successor is already in the frontier set, its path cost needs to be updated if lower than the previously computed one g(x): (Pessimistic) Estimate. Real cost-to-come BFS would stop here! G g(g) = 310 g(g) =
43 DIJKSTRA S ALGORITHM (1959) Shortest paths from a node s to all all other nodes Input: Graph G=(V,E) ( x s) dist[x] = + # dist[x] g(x) ~ distance from s (start) dist[s] = 0 C = # Closed set ~ Expanded nodes set O = V # Open set, ordered by dist[] ~ Our frontier set while O do u = extract_min(o) # expand node u C = C {u} # If u == goal: exit, since sp s u has been found foreach vertex v Successor(u) do dist[v] = min {dist[v], dist[u] + step_cost(u,v)} # Relaxation end do end do Time complexity: O( E + V log V ) using a Fibonacci heap for min queue 43
44 UNIFORM-COST SEARCH Algorithm UCS Complete? Optimal? Time Space Sorta Yes Θ b ;J K /M Θ b ;J K /M Completeness: If the cost of every step exceeds ε > 0 (and b is finite) there is only a finite number of expansions before the total path cost gets equal to the path cost of goal state. Instead, it could get into an infinite loop for ε = 0 or for costs infinitely decreasing On this branch, step-costs go as 2 -d 44
45 UNIFORM-COST SEARCH Algorithm UCS Complete? Optimal? Time Space Sorta Yes Θ b ;J K /M Θ b ;J K /M Optimality: (1) When a node n is selected for expansion, the shortest path to it has been found; otherwise another node m would be in the frontier (separation property) such that the path from start to n through m is shorter. But in this case m should have been selected before n for expansion (Contradiction!) (2) Because steps are non negative, paths never get shorter as nodes are added. UCS expands nodes (goals) in order of optimality and this is true also for the goal nodes 45
46 UNIFORM-COST SEARCH Algorithm UCS Complete? Optimal? Time Space Sorta Yes Θ b ;J K /M Θ b ;J K /M Optimality: Given completeness, the previous result ensures that when a goal node is selected for expansion the optimal path to it has been found Time complexity: If C is the cost of the optimal solution and ε is a lower bound on the step cost, the worst-case depth of the search tree is 1 + C /ε The complexity is Θ(b 5J; ) when step costs are uniform (in this case BFS is preferred) 46
47 DEPTH-FIRST SEARCH Strategy: Expand deepest unexpanded node Vertical diving + backtracking when reach a leaf Can be implemented by using a stack for the frontier (LIFO) Recursive implementation is also common (not use stack) 47
48 DFS FOR 8-PUZZLE Goal found after 19 node expansions and 32 goal checks 48
49 DFS FOR 8-PUZZLE Stack R 20L L
50 DEPTH-FIRST SEARCH Algorithm DFS Complete? Optimal? Time Space Finite Y No Θ b R Θ b m In a finite state space, which version of DFS is complete? 1. TREE SEARCH 2. GRAPH SEARCH 3. Both 4. Neither 50
51 DEPTH-FIRST SEARCH Graph search avoid loops, and in a finite space will eventually expand every node à Complete At almost zero cost, a check can be added to Tree search to avoid looping on the current path from start to the current node It makes DFS Tree search complete in finite space but doesn t avoid proliferation of redundant paths In infinite spaces, both versions fail for completeness if one diving branch is infinite Completeness: Clearly not in general x.. 51
52 DEPTH-FIRST SEARCH Algorithm DFS Complete? Optimal? Time Space In finite spaces No Θ b R Θ b m Optimality: Nope, if goal is on the branch being explored it will be reported as goal but there could be a better path Time complexity: Graph search is bounded by the size of the state space, it can generate up to the total number of states. Tree search is O b R, where m is the maximum depth of any node, can be much larger than state space! Space complexity: DFS tree search needs to store only a single path from the root to a leaf (m worst-case), along with unexpanded sibling nodes for each node on the path (b per node). Graph search is equivalent to BFS 52
53 DEPTH-FIRST BACKTRACKING SEARCH Recursive approach Only one successor is generated at a time When no further expansions are possible the search backtracks to the latest partially expanded node Each partially expanded node remembers which successor to generate next O(m) memory is needed It saves memory (and time) Particularly effective when successors are obtained by state modifications (CSPs) 53
54 DFS VS. BFS Algorithm DFS Algorithm BFS Complete? Optimal? Time Space In finite spaces No Θ b R Θ b m Complete? Optimal? Time Space Yes if Not really Θ b 5 Θ b 5 d, b finite Example: b=10, d=16=m BFS: nodes, 10 exabytes of memory if 1 node = 1 Kbytes DFS: = 160 nodes ~ 156 KB Depth-first tree search is the workhorse of many areas of AI 54
55 DFS VS. BFS Algorithm DFS Algorithm BFS Complete? Optimal? Time Space In finite spaces No Θ b R Θ b m Complete? Optimal? Time Space Yes if Not really Θ b 5 Θ b 5 d, b finite Performance depends on the structure of the search tree If goal is not far from root (shallow) BFS If the tree is deep (large m) and goal nodes are rare BFS If the branching factor b is large DFS If goal nodes are deep and frequent DFS If the search tree is infinite and b is finite BFS If the search tree is very deep DFS + limits for search depth If the search space is very large and depth is unknown IDS 55
56 DEPTH-LIMITED DF SEARCH Strategy: Expand deepest unexpanded node until a limit Run DFS with given depth limit l Additional source of incompleteness when l < d DFS is a special case of depth-limited for l = How to set l? Algorithm DL-DFS Complete? Optimal? Time Space If l > d No Θ b R Θ b m 56
57 ITERATIVE DEEPENING SEARCH (IDS) Limit = 0 A A Limit = 1 A A A A B C B C B C B C Strategy: Expand deepest unexpanded node until a depth limit l, and iteratively increase the limit l Limit = 2 Limit = 3 B A D E F G B A D E F G A C C A B C D E F G A B C D E F G A A B C D E F G A B C D E F G A A B C D E F G A B C D E F G A Run DFS with depth limit l = 1,2, until a goal is found B D E F G H I J K L M N O B A D E F G C C B C D E F G H I J K L M N O A B C D E F G B C D E F G H I J K L M N O A B C D E F G B C D E F G H I J K L M N O A B C D E F G Combines the best properties of BFS and DFS H I J K L M N O B A D E F G H I J K L M N O C H I J K L M N O A B C D E F G H I J K L M N O H I J K L M N O A B C D E F G H I J K L M N O H I J K L M N O A B C D E F G H I J K L M N O 57
58 ITERATIVE DEEPENING SEARCH Algorithm IDS Complete? Optimal? Time Space Yes No Θ b 5 Θ b d Completeness: Yes, for the same reason BFS is complete Optimal: Suffers from the same shortcoming of DFS Time complexity: Seems wasteful but most of the nodes are at the bottom level d, such that asymptotically is like BFS. For a fixed branching factor b, nodes at first level are generated d times, those at second level, d-1 times, and so on: d b + d 1 b C b 5 = Θ b 5 Space complexity: Enjoys the same properties as DFS 58
59 SUMMARY OF UNINFORMED Algorithm BFS UCS DFS IDS ALGORITHMS Complete? Optimal? Time Space Yes if d,b finite Not really Θ b 5 Θ b 5 Sorta Yes Θ b ;J K /M Θ b ;J K /M In finite spaces No Θ b R Θ b m Same as BFS No Θ b 5 Θ b d 59
60 SUMMARY (Model based, Deterministic) Automated planning, search problems Tree search and graph search meta-algorithms Performance metrics: completeness, optimality, time and space Uninformed search algorithms: BSF (problems with memory space) Bidirectional BFS (save in memory) UCS (~BFS for finding cost-optimal solutions) DFS (Tree search version has very low space requirements) Depth-limited DFS (avoid infinite diving) IDS (progressively increase depth limits) (Structure of search tree + Resources) vs. Algorithm 60
Chapter 3: Solving Problems by Searching
Chapter 3: Solving Problems by Searching Prepared by: Dr. Ziad Kobti 1 Problem-Solving Agent Reflex agent -> base its actions on a direct mapping from states to actions. Cannot operate well in large environments
More informationCS 4700: Foundations of Artificial Intelligence. Bart Selman. Search Techniques R&N: Chapter 3
CS 4700: Foundations of Artificial Intelligence Bart Selman Search Techniques R&N: Chapter 3 Outline Search: tree search and graph search Uninformed search: very briefly (covered before in other prerequisite
More informationSet 2: State-spaces and Uninformed Search. ICS 271 Fall 2015 Kalev Kask
Set 2: State-spaces and Uninformed Search ICS 271 Fall 2015 Kalev Kask You need to know State-space based problem formulation State space (graph) Search space Nodes vs. states Tree search vs graph search
More informationUninformed Search Methods
Uninformed Search Methods Search Algorithms Uninformed Blind search Breadth-first uniform first depth-first Iterative deepening depth-first Bidirectional Branch and Bound Informed Heuristic search Greedy
More informationAr#ficial)Intelligence!!
Introduc*on! Ar#ficial)Intelligence!! Roman Barták Department of Theoretical Computer Science and Mathematical Logic Problem Solving: Uninformed Search Simple reflex agent only transfers the current percept
More informationCAP 4630 Artificial Intelligence
CAP 4630 Artificial Intelligence Instructor: Sam Ganzfried sganzfri@cis.fiu.edu 1 http://www.ultimateaiclass.com/ https://moodle.cis.fiu.edu/ 2 Solving problems by search 3 8-puzzle 4 8-queens 5 Search
More informationCS 771 Artificial Intelligence. Problem Solving by Searching Uninformed search
CS 771 Artificial Intelligence Problem Solving by Searching Uninformed search Complete architectures for intelligence? Search? Solve the problem of what to do. Learning? Learn what to do. Logic and inference?
More informationChapter 3. A problem-solving agent is a kind of goal-based agent. It decide what to do by finding sequences of actions that lead to desirable states.
Chapter 3 A problem-solving agent is a kind of goal-based agent. It decide what to do by finding sequences of actions that lead to desirable states. A problem can be defined by four components : 1. The
More informationUninformed Search. Problem-solving agents. Tree search algorithms. Single-State Problems
Uninformed Search Problem-solving agents Tree search algorithms Single-State Problems Breadth-First Search Depth-First Search Limited-Depth Search Iterative Deepening Extensions Graph search algorithms
More informationSolving Problem by Searching. Chapter 3
Solving Problem by Searching Chapter 3 Outline Problem-solving agents Problem formulation Example problems Basic search algorithms blind search Heuristic search strategies Heuristic functions Problem-solving
More informationSearch Algorithms. Uninformed Blind search. Informed Heuristic search. Important concepts:
Uninformed Search Search Algorithms Uninformed Blind search Breadth-first uniform first depth-first Iterative deepening depth-first Bidirectional Branch and Bound Informed Heuristic search Greedy search,
More informationSearch I. slides from: Padhraic Smyth, Bryan Low, S. Russell and P. Norvig
Search I slides from: Padhraic Smyth, Bryan Low, S. Russell and P. Norvig Problem-Solving Agents Intelligent agents can solve problems by searching a state-space State-space Model the agent s model of
More informationARTIFICIAL INTELLIGENCE SOLVING PROBLEMS BY SEARCHING. Chapter 3
ARTIFICIAL INTELLIGENCE SOLVING PROBLEMS BY SEARCHING Chapter 3 1 PROBLEM SOLVING We want: To automatically solve a problem We need: A representation of the problem Algorithms that use some strategy to
More informationBasic Search. Fall Xin Yao. Artificial Intelligence: Basic Search
Basic Search Xin Yao Fall 2017 Fall 2017 Artificial Intelligence: Basic Search Xin Yao Outline Motivating Examples Problem Formulation From Searching to Search Tree Uninformed Search Methods Breadth-first
More informationArtificial Intelligence
Artificial Intelligence Search Marc Toussaint University of Stuttgart Winter 2015/16 (slides based on Stuart Russell s AI course) Outline Problem formulation & examples Basic search algorithms 2/100 Example:
More informationChapter3. Problem-Solving Agents. Problem Solving Agents (cont.) Well-defined Problems and Solutions. Example Problems.
Problem-Solving Agents Chapter3 Solving Problems by Searching Reflex agents cannot work well in those environments - state/action mapping too large - take too long to learn Problem-solving agent - is one
More informationARTIFICIAL INTELLIGENCE. Pathfinding and search
INFOB2KI 2017-2018 Utrecht University The Netherlands ARTIFICIAL INTELLIGENCE Pathfinding and search Lecturer: Silja Renooij These slides are part of the INFOB2KI Course Notes available from www.cs.uu.nl/docs/vakken/b2ki/schema.html
More informationArtificial Intelligence
Artificial Intelligence CSC348 Unit 3: Problem Solving and Search Syedur Rahman Lecturer, CSE Department North South University syedur.rahman@wolfson.oxon.org Artificial Intelligence: Lecture Notes The
More information521495A: Artificial Intelligence
521495A: Artificial Intelligence Search Lectured by Abdenour Hadid Associate Professor, CMVS, University of Oulu Slides adopted from http://ai.berkeley.edu Agent An agent is an entity that perceives the
More informationSolving problems by searching
Solving problems by searching Chapter 3 Systems 1 Outline Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms Systems 2 Problem-solving agents Systems 3 Example:
More informationSolving problems by searching
Solving problems by searching Chapter 3 Some slide credits to Hwee Tou Ng (Singapore) Outline Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms Heuristics
More informationCS 380: Artificial Intelligence Lecture #3
CS 380: Artificial Intelligence Lecture #3 William Regli Outline Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms 1 Problem-solving agents Example: Romania
More informationGraphs vs trees up front; use grid too; discuss for BFS, DFS, IDS, UCS Cut back on A* optimality detail; a bit more on importance of heuristics,
Graphs vs trees up front; use grid too; discuss for BFS, DFS, IDS, UCS Cut back on A* optimality detail; a bit more on importance of heuristics, performance data Pattern DBs? General Tree Search function
More informationARTIFICIAL INTELLIGENCE (CSC9YE ) LECTURES 2 AND 3: PROBLEM SOLVING
ARTIFICIAL INTELLIGENCE (CSC9YE ) LECTURES 2 AND 3: PROBLEM SOLVING BY SEARCH Gabriela Ochoa http://www.cs.stir.ac.uk/~goc/ OUTLINE Problem solving by searching Problem formulation Example problems Search
More informationSolving Problems by Searching
INF5390 Kunstig intelligens Sony Vaio VPC-Z12 Solving Problems by Searching Roar Fjellheim Outline Problem-solving agents Example problems Search programs Uninformed search Informed search Summary AIMA
More informationChapter 3 Solving problems by searching
1 Chapter 3 Solving problems by searching CS 461 Artificial Intelligence Pinar Duygulu Bilkent University, Slides are mostly adapted from AIMA and MIT Open Courseware 2 Introduction Simple-reflex agents
More informationChapter 2. Blind Search 8/19/2017
Chapter 2 1 8/19/2017 Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms 8/19/2017 2 8/19/2017 3 On holiday in Romania; currently in Arad. Flight leaves tomorrow
More informationSolving Problems by Searching
INF5390 Kunstig intelligens Solving Problems by Searching Roar Fjellheim Outline Problem-solving agents Example problems Search programs Uninformed search Informed search Summary AIMA Chapter 3: Solving
More informationCS 8520: Artificial Intelligence
CS 8520: Artificial Intelligence Solving Problems by Searching Paula Matuszek Spring, 2013 Slides based on Hwee Tou Ng, aima.eecs.berkeley.edu/slides-ppt, which are in turn based on Russell, aima.eecs.berkeley.edu/slides-pdf.
More informationUninformed search strategies (Section 3.4) Source: Fotolia
Uninformed search strategies (Section 3.4) Source: Fotolia Uninformed search strategies A search strategy is defined by picking the order of node expansion Uninformed search strategies use only the information
More information3 SOLVING PROBLEMS BY SEARCHING
48 3 SOLVING PROBLEMS BY SEARCHING A goal-based agent aims at solving problems by performing actions that lead to desirable states Let us first consider the uninformed situation in which the agent is not
More informationSolving problems by searching
Solving problems by searching 1 C H A P T E R 3 Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms Outline 2 Problem-solving agents 3 Note: this is offline
More informationCPS 170: Artificial Intelligence Search
CPS 170: Artificial Intelligence http://www.cs.duke.edu/courses/spring09/cps170/ Search Instructor: Vincent Conitzer Search We have some actions that can change the state of the world Change resulting
More informationHW#1 due today. HW#2 due Monday, 9/09/13, in class Continue reading Chapter 3
9-04-2013 Uninformed (blind) search algorithms Breadth-First Search (BFS) Uniform-Cost Search Depth-First Search (DFS) Depth-Limited Search Iterative Deepening Best-First Search HW#1 due today HW#2 due
More informationPengju XJTU 2016
Introduction to AI Chapter03 Solving Problems by Uninformed Searching(3.1~3.4) Pengju Ren@IAIR Outline Problem-solving agents Problem types Problem formulation Search on Trees and Graphs Uninformed algorithms
More informationLecture 2: Fun with Search. Rachel Greenstadt CS 510, October 5, 2017
Lecture 2: Fun with Search Rachel Greenstadt CS 510, October 5, 2017 Reminder! Project pre-proposals due tonight Overview Uninformed search BFS, DFS, Uniform-Cost, Graph-Search Informed search Heuristics,
More informationAGENTS AND ENVIRONMENTS. What is AI in reality?
AGENTS AND ENVIRONMENTS What is AI in reality? AI is our attempt to create a machine that thinks (or acts) humanly (or rationally) Think like a human Cognitive Modeling Think rationally Logic-based Systems
More informationEE562 ARTIFICIAL INTELLIGENCE FOR ENGINEERS
EE562 ARTIFICIAL INTELLIGENCE FOR ENGINEERS Lecture 3, 4/6/2005 University of Washington, Department of Electrical Engineering Spring 2005 Instructor: Professor Jeff A. Bilmes 4/6/2005 EE562 1 Today: Basic
More informationSolving Problems by Searching
Solving Problems by Searching Berlin Chen 2004 Reference: 1. S. Russell and P. Norvig. Artificial Intelligence: A Modern Approach. Chapter 3 1 Introduction Problem-Solving Agents vs. Reflex Agents Problem-solving
More informationProblem solving and search
Problem solving and search Chapter 3 Chapter 3 1 Problem formulation & examples Basic search algorithms Outline Chapter 3 2 On holiday in Romania; currently in Arad. Flight leaves tomorrow from Bucharest
More informationCS 5522: Artificial Intelligence II
CS 5522: Artificial Intelligence II Search Algorithms Instructor: Wei Xu Ohio State University [These slides were adapted from CS188 Intro to AI at UC Berkeley.] Today Agents that Plan Ahead Search Problems
More informationCS 4100 // artificial intelligence
CS 4100 // artificial intelligence instructor: byron wallace Search I Attribution: many of these slides are modified versions of those distributed with the UC Berkeley CS188 materials Thanks to John DeNero
More informationArtificial Intelligence Problem Solving and Uninformed Search
Artificial Intelligence Problem Solving and Uninformed Search Maurizio Martelli, Viviana Mascardi {martelli, mascardi}@disi.unige.it University of Genoa Department of Computer and Information Science AI,
More informationProblem Solving and Search
Artificial Intelligence Problem Solving and Search Dae-Won Kim School of Computer Science & Engineering Chung-Ang University Outline Problem-solving agents Problem types Problem formulation Example problems
More informationAnnouncements. Project 0: Python Tutorial Due last night
Announcements Project 0: Python Tutorial Due last night HW1 officially released today, but a few people have already started on it Due Monday 2/6 at 11:59 pm P1: Search not officially out, but some have
More informationCSC 2114: Artificial Intelligence Search
CSC 2114: Artificial Intelligence Search Ernest Mwebaze emwebaze@cit.ac.ug Office: Block A : 3 rd Floor [Slide Credit Dan Klein and Pieter Abbeel for CS188 Intro to AI at UC Berkeley. Reference materials
More informationSolving problems by searching
Solving problems by searching Chapter 3 CS 2710 1 Outline Problem-solving agents Problem formulation Example problems Basic search algorithms CS 2710 - Blind Search 2 1 Goal-based Agents Agents that take
More informationAI: problem solving and search
: problem solving and search Stefano De Luca Slides mainly by Tom Lenaerts Outline Problem-solving agents A kind of goal-based agent Problem types Single state (fully observable) Search with partial information
More informationUninformed Search (Ch )
Uninformed Search (Ch. 3-3.4) Announcements First homework will be posted tonight (due next Wednesday at 11:55 pm) Review We use words that have a general English definition in a technical sense nd Rational=choose
More informationSolving problems by searching. Chapter 3
Solving problems by searching Chapter 3 Outline Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms 2 Example: Romania On holiday in Romania; currently in
More informationOutline. Solving problems by searching. Problem-solving agents. Example: Romania
Outline Solving problems by searching Chapter 3 Problem-solving agents Problem types Problem formulation Example problems Basic search algorithms Systems 1 Systems 2 Problem-solving agents Example: Romania
More informationMultiagent Systems Problem Solving and Uninformed Search
Multiagent Systems Problem Solving and Uninformed Search Viviana Mascardi viviana.mascardi@unige.it MAS, University of Genoa, DIBRIS Classical AI 1 / 36 Disclaimer This presentation may contain material
More informationProblem solving and search
Problem solving and search Chapter 3 Chapter 3 1 Outline Problem-solving agents Problem types Problem formulation Example problems Uninformed search algorithms Informed search algorithms Chapter 3 2 Restricted
More informationAGENTS AND ENVIRONMENTS. What is AI in reality?
AGENTS AND ENVIRONMENTS What is AI in reality? AI is our attempt to create a machine that thinks (or acts) humanly (or rationally) Think like a human Cognitive Modeling Think rationally Logic-based Systems
More informationUninformed (also called blind) search algorithms
Uninformed (also called blind) search algorithms First Lecture Today (Thu 30 Jun) Read Chapters 18.6.1-2, 20.3.1 Second Lecture Today (Thu 30 Jun) Read Chapter 3.1-3.4 Next Lecture (Tue 5 Jul) Chapters
More information521495A: Artificial Intelligence
521495A: Artificial Intelligence Informed Search Lectured by Abdenour Hadid Adjunct Professor, CMVS, University of Oulu Slides adopted from http://ai.berkeley.edu Today Informed Search Heuristics Greedy
More informationSolving Problems by Searching
Solving Problems by Searching Agents, Goal-Based Agents, Problem-Solving Agents Search Problems Blind Search Strategies Agents sensors environment percepts actions? agent effectors Definition. An agent
More informationSolving Problems by Searching
Solving Problems by Searching Agents, Goal-Based Agents, Problem-Solving Agents Search Problems Blind Search Strategies Agents sensors environment percepts actions? agent effectors Definition. An agent
More informationITCS 6150 Intelligent Systems. Lecture 3 Uninformed Searches
ITCS 6150 Intelligent Systems Lecture 3 Uninformed Searches Outline Problem Solving Agents Restricted form of general agent Problem Types Fully vs. partially observable, deterministic vs. stochastic Problem
More informationSolving Problems by Searching
Solving Problems by Searching CS486/686 University of Waterloo Sept 11, 2008 1 Outline Problem solving agents and search Examples Properties of search algorithms Uninformed search Breadth first Depth first
More informationS A E RC R H C I H NG N G IN N S T S A T T A E E G R G A R PH P S
LECTURE 2 SEARCHING IN STATE GRAPHS Introduction Idea: Problem Solving as Search Basic formalism as State-Space Graph Graph explored by Tree Search Different algorithms to explore the graph Slides mainly
More informationWeek 3: Path Search. COMP9414/ 9814/ 3411: Artificial Intelligence. Motivation. Example: Romania. Romania Street Map. Russell & Norvig, Chapter 3.
COMP9414/9814/3411 17s1 Search 1 COMP9414/ 9814/ 3411: Artificial Intelligence Week 3: Path Search Russell & Norvig, Chapter 3. Motivation Reactive and Model-Based Agents choose their actions based only
More informationUninformed Search. CS171, Winter 2018 Introduction to Artificial Intelligence Prof. Richard Lathrop. Reading: R&N
Uninformed Search CS171, Winter 2018 Introduction to Artificial Intelligence Prof. Richard Lathrop Reading: R&N 3.1-3.4 Uninformed search strategies Uninformed (blind): You have no clue whether one non-goal
More informationPengju
Introduction to AI Chapter03 Solving Problems by Uninformed Searching(3.1~3.4) Pengju Ren@IAIR Outline Problem-solving agents Problem types Problem formulation Search on Trees and Graphs Uninformed algorithms
More informationCS486/686 Lecture Slides (c) 2015 P.Poupart
1 2 Solving Problems by Searching [RN2] Sec 3.1-3.5 [RN3] Sec 3.1-3.4 CS486/686 University of Waterloo Lecture 2: May 7, 2015 3 Outline Problem solving agents and search Examples Properties of search algorithms
More informationSolving problems by searching
Solving problems by searching CE417: Introduction to Artificial Intelligence Sharif University of Technology Spring 2017 Soleymani Artificial Intelligence: A Modern Approach, Chapter 3 Outline Problem-solving
More informationSolving problems by searching
Solving problems by searching CE417: Introduction to Artificial Intelligence Sharif University of Technology Spring 2014 Soleymani Artificial Intelligence: A Modern Approach, Chapter 3 Outline Problem-solving
More informationClass Overview. Introduction to Artificial Intelligence COMP 3501 / COMP Lecture 2. Problem Solving Agents. Problem Solving Agents: Assumptions
Class Overview COMP 3501 / COMP 4704-4 Lecture 2 Prof. JGH 318 Problem Solving Agents Problem Solving Agents: Assumptions Requires a goal Assume world is: Requires actions Observable What actions? Discrete
More informationProblem Solving as Search. CMPSCI 383 September 15, 2011
Problem Solving as Search CMPSCI 383 September 15, 2011 1 Today s lecture Problem-solving as search Uninformed search methods Problem abstraction Bold Claim: Many problems faced by intelligent agents,
More informationRoute planning / Search Movement Group behavior Decision making
Game AI Where is the AI Route planning / Search Movement Group behavior Decision making General Search Algorithm Design Keep a pair of set of states: One, the set of states to explore, called the open
More informationCS486/686 Lecture Slides (c) 2014 P.Poupart
1 2 1 Solving Problems by Searching [RN2] Sec 3.1-3.5 [RN3] Sec 3.1-3.4 CS486/686 University of Waterloo Lecture 2: January 9, 2014 3 Outline Problem solving agents and search Examples Properties of search
More informationGoal-Based Agents Problem solving as search. Outline
Goal-Based Agents Problem solving as search Vasant Honavar Bioinformatics and Computational Biology Program Center for Computational Intelligence, Learning, & Discovery honavar@cs.iastate.edu www.cs.iastate.edu/~honavar/
More informationSolving Problems by Searching
Solving Problems by Searching 1 Terminology State State Space Initial State Goal Test Action Step Cost Path Cost State Change Function State-Space Search 2 Formal State-Space Model Problem = (S, s, A,
More informationArtificial Intelligence
Artificial Intelligence Dr Ahmed Rafat Abas Computer Science Dept, Faculty of Computers and Informatics, Zagazig University arabas@zu.edu.eg http://www.arsaliem.faculty.zu.edu.eg/ Solving problems by searching
More informationInformed Search CS457 David Kauchak Fall 2011
Admin Informed Search CS57 David Kauchak Fall 011 Some material used from : Sara Owsley Sood and others Q3 mean: 6. median: 7 Final projects proposals looked pretty good start working plan out exactly
More informationInformed (Heuristic) Search. Idea: be smart about what paths to try.
Informed (Heuristic) Search Idea: be smart about what paths to try. 1 Blind Search vs. Informed Search What s the difference? How do we formally specify this? A node is selected for expansion based on
More informationSearch EECS 395/495 Intro to Artificial Intelligence
Search EECS 395/495 Intro to Artificial Intelligence Doug Downey (slides from Oren Etzioni, based on Stuart Russell, Dan Weld, Henry Kautz, and others) What is Search? Search is a class of techniques for
More informationSearch EECS 348 Intro to Artificial Intelligence
Search EECS 348 Intro to Artificial Intelligence (slides from Oren Etzioni, based on Stuart Russell, Dan Weld, Henry Kautz, and others) What is Search? Search is a class of techniques for systematically
More informationCS 151: Intelligent Agents, Problem Formulation and Search
CS 151: Intelligent Agents, Problem Formulation and Search How do we make a computer "smart?" Computer, clean the house! Um OK?? This one's got no chance How do we represent this problem? Hmmm where to
More informationClass Overview. Introduction to Artificial Intelligence COMP 3501 / COMP Lecture 2: Search. Problem Solving Agents
Class Overview COMP 3501 / COMP 4704-4 Lecture 2: Search Prof. 1 2 Problem Solving Agents Problem Solving Agents: Assumptions Requires a goal Assume world is: Requires actions Observable What actions?
More informationCS540 Uninformed Search
CS540 Uninformed Search Xiaojin Zhu jerryzhu@cs.wisc.edu Computer Sciences Department University of Wisconsin, Madison slide 1 Main messages Many AI problems can be formulated as search. Iterative deepening
More informationSearch : Lecture 2. September 9, 2003
Search 6.825: Lecture 2 September 9, 2003 1 Problem-Solving Problems When your environment can be effectively modeled as having discrete states and actions deterministic, known world dynamics known initial
More informationLecture 3. Uninformed Search
Lecture 3 Uninformed Search 1 Uninformed search strategies Uninformed: While searching you have no clue whether one non-goal state is better than any other. Your search is blind. You don t know if your
More informationSolving Problems: Blind Search
Solving Problems: Blind Search Instructor: B. John Oommen Chancellor s Professor Fellow: IEEE ; Fellow: IAPR School of Computer Science, Carleton University, Canada The primary source of these notes are
More information4. Solving Problems by Searching
COMP9414/9814/3411 15s1 Search 1 COMP9414/ 9814/ 3411: Artificial Intelligence 4. Solving Problems by Searching Russell & Norvig, Chapter 3. Motivation Reactive and Model-Based Agents choose their actions
More informationLecture 4: Search 3. Victor R. Lesser. CMPSCI 683 Fall 2010
Lecture 4: Search 3 Victor R. Lesser CMPSCI 683 Fall 2010 First Homework 1 st Programming Assignment 2 separate parts (homeworks) First part due on (9/27) at 5pm Second part due on 10/13 at 5pm Send homework
More informationSolving Problems by Searching (Blindly)
Solving Problems by Searching (Blindly) R&N: Chap. 3 (many of these slides borrowed from Stanford s AI Class) Problem Solving Agents Decide what to do by finding a sequence of actions that lead to desirable
More informationProblem Solving & Heuristic Search
190.08 Artificial 2016-Spring Problem Solving & Heuristic Search Byoung-Tak Zhang School of Computer Science and Engineering Seoul National University 190.08 Artificial (2016-Spring) http://www.cs.duke.edu/courses/fall08/cps270/
More informationInformed Search Lecture 5
Lecture 5 How can we exploit problem-specific knowledge to find solutions more efficiently? Where does this knowledge come from and under what conditions is it useful? 1 Agenda Review: Uninformed Search
More informationA.I.: Informed Search Algorithms. Chapter III: Part Deux
A.I.: Informed Search Algorithms Chapter III: Part Deux Best-first search Greedy best-first search A * search Heuristics Outline Overview Informed Search: uses problem-specific knowledge. General approach:
More informationCS510 \ Lecture Ariel Stolerman
CS510 \ Lecture02 2012-10-03 1 Ariel Stolerman Midterm Evan will email about that after the lecture, at least 2 lectures from now. The exam will be given in a regular PDF (not an online form). We will
More informationArtificial Intelligence Uninformed search
Artificial Intelligence Uninformed search A.I. Uninformed search 1 The symbols&search hypothesis for AI Problem-solving agents A kind of goal-based agent Problem types Single state (fully observable) Search
More informationPEAS: Medical diagnosis system
PEAS: Medical diagnosis system Performance measure Patient health, cost, reputation Environment Patients, medical staff, insurers, courts Actuators Screen display, email Sensors Keyboard/mouse Environment
More informationUninformed Search. Reading: Chapter 4 (Tuesday, 2/5) HW#1 due next Tuesday
Uninformed Search Reading: Chapter 4 (Tuesday, 2/5) HW#1 due next Tuesday 1 Uninformed Search through the space of possible solutions Use no knowledge about which path is likely to be best Exception: uniform
More informationSearch. (Textbook Chpt ) Computer Science cpsc322, Lecture 2. May, 10, CPSC 322, Lecture 2 Slide 1
Search Computer Science cpsc322, Lecture 2 (Textbook Chpt 3.0-3.4) May, 10, 2012 CPSC 322, Lecture 2 Slide 1 Colored Cards You need to have 4 colored index cards Come and get them from me if you still
More informationBlind (Uninformed) Search (Where we systematically explore alternatives)
Blind (Uninformed) Search (Where we systematically explore alternatives) R&N: Chap. 3, Sect. 3.3 5 Slides from Jean-Claude Latombe at Stanford University (used with permission) Simple Problem-Solving-Agent
More informationKI-Programmierung. Basic Search Algorithms
KI-Programmierung Basic Search Algorithms Bernhard Beckert UNIVERSITÄT KOBLENZ-LANDAU Winter Term 2007/2008 B. Beckert: KI-Programmierung p.1 Example: Travelling in Romania Scenario On holiday in Romania;
More informationCS242 Uninformed Search
CS242 Uninformed Search Lecturer: Ji Liu Thanks for Jerry Zhu's slides slide 1 Main messages Many (classic) AI problems can be formulated as search. Iterative deepening is good when you don t know much.
More informationAr#ficial)Intelligence!!
Introduc*on! Ar#ficial)Intelligence!! Roman Barták Department of Theoretical Computer Science and Mathematical Logic Uninformed (blind) search algorithms can find an (optimal) solution to the problem,
More informationInformed search methods
Informed search methods Tuomas Sandholm Computer Science Department Carnegie Mellon University Read Section 3.5-3.7 of Russell and Norvig Informed Search Methods Heuristic = to find, to discover Heuristic
More information