Real-Time Heuristic Search: First

Existing heuristic search algorithms are not applicable to real-time applications because they cannot commit to a move before an entire solution is found. We present a special case of minimax lookahead search to handle this problem, and an analog of alpha-beta pruning that significantly improves the efficiency of the algorithm. In addition, we present a new algorithm, called Real-Time-A*, for searching when actions must actually be executed, as opposed to merely simulated. Finally, we examine the nature of the tradeoff between computation and execution cost. EIeuristic search is a fundamental problem-solving method in artificial intelligence. l?or most AI problems, the sequence of steps required for solution is not known a priori but must be determined by a systematic trial-and-error exploration of alternatives. All that is required to formulate a search problem is a set of states, a set of operators that map states to states, an initial state, and a set of goal states. The task typically is to find a lowest cost sequence of operators that map the initial state to a goal state. The complexity of search algorithms is greatly reduced by the use of a heuristic evaluation function, often without sacrificing solution optima&y. A heuristic is a function that is relatively cheap to compute and estimates the cost of the cheapest path from a given state to a goal state. Common examples in the AI literature of search problems are the Eight Puzzle and its larger relative the Fifteen Puzzle. The Eight Puzzle consists of a 3x3 square frame containing 8 numbered square tiles and an empty position called the “blank”. The legal operators slide any tile horizontally or vertically adjacent to the blank into the blank position. The task is to rearrange the tiles from some random initial configuration into a particular desired goal configuration. A common heuristic function for this problem is called Manhattan Distance. It is computed by counting, for each tile not ‘in its goal position, the number of moves along the grid it is away from its goal position, and summing these values over all tiles, excluding the blank. A real-world example is the task of autonomous navigation in a network of roads, or arbitrary terrain, from an initial location to a desired goal location. The problem is typically to find a shorted path between the initial and goal states. A typical heuristic evaluation function for this problem is the air-line distance from a given location to the