The Stanford Research Institute Problem Solver, or STRIPS, is a foundational automated planning system originally developed at the Stanford Research Institute (now SRI International) in the early 1970s. STRIPS was created by Richard Fikes and Nils Nilsson as part of research into enabling robots and AI agents to plan sequences of actions in complex environments. STRIPS introduced a formal language and approach for representing actions, world states, and their effects, and it has had an enormous influence on the field of artificial intelligence, especially in planning and robotics.
At its core, STRIPS represents the world as a set of facts, or predicates, that describe the current state. For example, you might have predicates like “at(robot, roomA)” or “box(box1)”. Actions are defined as operators, each of which has a set of preconditions that must be true for the action to be executed and a set of effects that describe how the world changes when the action is performed. For instance, a “move” action might require that the robot is in roomA and that the door to roomB is open. The effect of executing the action would be to update the fact that the robot is now in roomB.
The STRIPS planner takes as input an initial state, a goal condition (what facts need to be true at the end), and a set of possible actions. Using these, it searches for a sequence of actions that transforms the initial state into a state that satisfies the goal. This search can be guided by heuristics to make it more efficient, but in its simplest form, STRIPS performs a kind of logical search through the possible states.
One of the key contributions of STRIPS is its action representation, which is compact and allows for efficient reasoning. The STRIPS formalism is still widely taught in AI courses as a clear way to introduce students to automated planning. Many modern planning systems, including those used in robotics and logistics, are direct descendants of the STRIPS approach or use its representation as a foundation. STRIPS also introduced the concept of separating the description of actions from the specifics of a problem instance, making it possible to use the same set of actions across many different planning problems.
Despite its age, STRIPS remains relevant because it captures the core challenges of planning: representing states, actions, and goals, and efficiently searching for solutions. Its influence extends to areas like reinforcement learning, symbolic AI, and even natural language understanding, wherever structured representations of actions and state transitions are needed.
