ABSTRACT
By design, an asynchronous, distributed, algorithm for a transportation system must necessarily reflect the highest, meta-level purpose or intent of the system. The algorithm manifests itself in the behavior of every constituent unit. It will hold the potential of exploiting the maximal parallelism inherent in the system. Furthermore, local computations must be maximized while minimizing the communications between the entities, thereby implying high throughput, robustness, and scalability. The key properties of such algorithms are as follows:
a. Identification and definition of entities: From the perspective of control, coordination, and resource allocation algorithm, entities constitute the basic decision-making units and define the resolution of the decision behavior of the transportation system. From the perspective of the physical system, entities correspond to its natural, constituent elements and include the resources and the units to be transported.An entity must be self-contained, that is, its behavior, under every possible scenario, is completely defined within itself. Every entity exists independent of all other entities and, therefore, its behavior is known only to itself. Unless the entity shares its behavior with a different entity, no one has knowledge of its unique behavior. Conceivably, an entity will interact with other entities. Under such conditions, its behavior must include the scope and nature of the interactions between itself and other entities.
