One goal of artificial intelligence is to need to replace human work, which is required to make a series of movements to the goal, that is, you can planning! Planning can actually be seen as a solution process: there is known (initial state), there is a corresponding logic required to find the problem (a set of system actions), so as long as the search finds the final state The path is a feasible plan. But how do you do it? We can use the corresponding theorem when we solve the problem. It also has the corresponding law. English also has a grammar. They have their own logic to describe the knowledge in the field, and artificial intelligence should go to "reasoning". Your own logic (everyone can find the book of this person's intelligence) about the introduction of logical systems, like proposition logic, predicate logic, and some non-classic logic, etc.). It can be reasoned on the basis of these logic. For example, suppose I have a robot, his task is to help me buy daily necessities (milk and banana), we use first-order proposition logic to describe this problem, the robot has two states: AT (x) indicates its location, Have (x) indicates all of its goods, and to complete the task it has action: Go (x, y) from x to y, prerequisites are AT (x), execute AT (Y), Buy (x, y) From Y to purchase the product x, prerequisites AT (x), execute the result, ", and assume that there is a supermarket supermarket selling the corresponding item. Then, its initial state is AT (Home), the target status Have (Milk), Have (banana). In this way, the robot with the corresponding search algorithm can complete the task :)
