Information Processing Device Optimizing Vehicle Routing Efficiency

The patent segments the route search process into two distinct phases: (1) generating multiple candidate routes that satisfy basic conditions, and (2) selecting the optimum route from candidates using an optimization device. This segmentation allows the system to handle large numbers of nodes efficiently by separating route generation from optimization, reducing overall processing time while maintaining accuracy.

The patent applies preliminary action by pre-generating multiple route candidates that satisfy basic conditions before performing optimization. The generation unit creates routes in advance that meet constraints such as time windows and vehicle capacities, so that the subsequent optimization process only needs to evaluate these pre-filtered candidates, significantly reducing the computational burden when the number of nodes increases.

The patent segments constraint satisfaction into two layers: basic constraints are enforced during route generation by the generation unit, while more complex optimization constraints are handled separately by the optimization device. This segmentation allows each component to focus on specific constraint types, reducing overall system complexity while ensuring all conditions are met.

The patent introduces an intermediary optimization device that acts as a mediator between the generation unit and the final route selection. This intermediary receives multiple candidate routes satisfying basic conditions and applies additional optimization criteria, effectively managing the complexity of multiple constraints without requiring the generation unit to handle all constraint types simultaneously.

The patent applies partial action by generating multiple candidate routes that are sufficient for finding the optimum without exhaustively evaluating all possible routes. The generation unit creates a set of routes that includes the optimum route with high probability, and the optimization device efficiently identifies it from this subset, achieving near-optimal results with significantly reduced computational effort compared to exhaustive search.

The patent uses preliminary action by pre-generating multiple feasible route candidates before optimization. This preliminary set of routes is carefully constructed to include the optimum route while filtering out clearly suboptimal options, so that the subsequent optimization process operates on a manageable subset rather than all possible routes, improving both accuracy and efficiency.