Implementation of artificial intelligence planning for seaport logistic operations using linear programming and three-valued logic system

This paper presents a method for translating PDDL-based automated planning into Linear Programming (LP) for controlling selected operations of an automated container terminal. We develop a unified mapping from PDDL conditions and operators to LP variables and constraints that supports conditional planning and uncertainty in the initial state. Uncertainty is represented using Kleene's three-valued logic (T = 1, F = 0, U = 0.5). New sensory operators add constraints that resolve unknown condition states to known ones before a decision is made. The approach is evaluated on a setup derived from an automated terminal in Shanghai, across tasks of increasing scale—from a two-step plan to larger sequences with up to 14 operators across 13 steps. Results show that the LP formulation computes plan within milliseconds to seconds and scales linearly with plan length, whereas classical PDDL solving (and mixed-integer variants) requires longer runtimes. The proposed LP-based framework enables fast, predictable plan generation with sensory operators, making it suitable for real-time decision support in Industry 4.0 logistics.