Problem Solving Thursday 23 March 2023 08:48 ? Engaging in Problem solving is ... • Cyclical: enact steps that occur in a loop 1. Define 2. Brainstorm solutions 3. Pick a solution 4. Implement the solution 5. Review the solution · Recursive: repeat the cycle as many times as necessary · Applicable: apply successful cycles to new problems ? Well defined problems · Goal directedness Problems with a well defined goal state and set task constraint such that there are clear steps Ex: solving sudoku; it has set rules that we need to follow · III defined problems · Ambiguous situations that: - Have limitations for how to solve the problem - You must create your own path - Has multiple solutions · Social problem solving is a form of ill defined problem solving . They carry a load according to their content We have to look at constraints -> solve anagram Well-defined (Constraints) Ill-defined (No constraints) think
only of music Constrains one to 1 Can you make a type of music with ZJAZ? Can you make a word with ZJAZ? Greater activity in the right lateral PFC for ill defined anagrams -> it carries a greater cognitive load Cognitive load: amount of information held in mind at one time. ? Moravec's paradox Artificial intelligence can solve well defined problems well but not ill-defined problems and simple skills. Because we use algorithms to define AI, it has deep neural networks that work well with certainty. ? Problem solving algorithms Strategies to move through a 'problem space' = a representation that includes: - Initial and goal states - Intermediate paths and operators: actions to change between states - Task constraints Tower of Hanoi 1. Initial and goals states 2. Intermediate states and operators 1 START & 21 20 19 Tf- 27 TIF 22 2 3 4 IFT 4. Task constraints 17 18 26 5 23 € 16 IFT 25 24 IT 7 |& GOAL 15 14 13 12 IF 11 10 9 8 Task constraints: move one at a time, bigger can't go on bigger
? Different approaches • Brute force approach: Systematic algorithm that represents all the possible steps from the problem to goal state -> it is guaranteed to find a solution but is inefficient -> Combinatorial explosion: we are computing too many alternatives · Narrowing down search: heuristics: Strategies are used to selected moves in a problem space -> it helps avoid the combinatorial explosion Hill climbing strategy Means end analysis ? Hill climbing strategy Difference reduction strategy. We select the operation that brings us closer to the goal without examining the whole problem space. It can lead to a local maxima: false outcome that we mistake as final goal Doesn't always work because some problems need us to move away from the goal to solve it ? Hobbits and orcs problem 3 hobbits and 3 orcs with two seats boat, want to go on the other side of the river -> if more orcs, they die ->