search
Color the map, meet the constraints
It is the era when scheduling and assignment problems flood into AI: timetables where no two exams clash, frequency assignments where neighbors must differ, layouts where every constraint must hold at once. Cast the problem abstractly — variables, domains of possible values, and constraints forbidding certain combinations — and search for an assignment satisfying them all. Blind backtracking drowns; the art is to propagate constraints so that fixing one variable prunes the domains of others before you ever guess them, detecting dead ends early and ordering choices to fail fast. Get it wrong and you explore a combinatorial swamp that a little propagation would have killed instantly, or accept a schedule violating a constraint no one checked — constraint satisfaction is the backbone of planning, and propagation makes it tractable.
Who this problem belongs to
The two figures whose methods fit it best, out of 68 in contention.
McCarthy's foundational vision for AI as symbolic reasoning over formally represented knowledge is the direct ancestor of constraint satisfaction as a paradigm: represent a problem as variables, domains, and logical constraints, then search systematically for an assignment satisfying them all. His LISP gave the field the practical language in which the first constraint-propagation and backtracking systems were actually written, and his work on circumscription and formalized common-sense reasoning shows deep, career-long comfort with exactly this kind of declarative problem representation. He did not personally invent arc consistency or the specific propagation algorithms modern CSP solvers use, which were developed by later researchers, but the entire conceptual framework this problem asks students to apply is substantially his intellectual property, keeping him near the top of this card.
Newell's Logic Theorist and General Problem Solver, built with Simon, were the first programs to search a symbolic space of partial solutions by systematically trying, testing, and backtracking from dead ends — precisely the backtracking-search paradigm at the heart of constraint satisfaction, developed years before CSP was named as a distinct subfield. His means-ends analysis, comparing the current state to the goal and choosing operators that reduce the difference, anticipates constraint propagation's logic of using known values to prune the domains of unknowns. His cognitive-architecture framing of problem-solving as heuristic search through a state space maps almost exactly onto this problem's scheduling and layout examples. The specific propagation algorithms (arc consistency, forward checking) came later from other researchers, which is the only real gap between his work and full authorship here.
Fought here
In the mind map
The same ideas, as concepts rather than history — in John's ML knowledge map.
68 figures are scored on this problem. Draw it in a battle to see where you land.