One page per recurring comprehension, in exam order. Each page states its rules, works a real past-paper question with a step-by-step animation, drills more real questions the same way, and ends with a checklist. Start at Search and go down the list.
Topics
Search algorithms: BFS, DFS, Best-First, A*, Branch and Bound on the river map; water jugs; DFID and wA* theory; admissibility test.
TSP Branch and Bound: frozen refinement trees; refined order, optimal node and cost, city count, tour from A. Five real trees, each animated.
Games: minimax, strategy and best strategy, alpha-beta pruning, SSS* solved leaves, eval assignment, mid-flight N cutoffs.
AO*: AND-OR graphs, the expansion loop, S-value tracking, admissibility verdicts. Four real graphs, each animated.
Rete net: conflict set, specificity, recency; variable-consistency and negative-pattern traps.
Automated planning: applicable (FSSP) and relevant (BSSP) actions, Layer 1 action and proposition mutexes.
CSP: modulo relations, arc consistency by revise, path-consistency witnesses, reading off the solution.
GraphPlan: mid-flight layers, applicability, action and proposition mutexes, planning-theory keys.
Logic goal trees: backward-chaining traces, leftmost-first expansion, backtracking on failure.
Diagnosis and revision: conflicts, hitting sets, Occam's razor; one-table revision of every topic above.