The shared data portions of the interface allows scheduling decisions to affect future planning decisions, including those decisions made when replanning upon scheduler failure. When scheduling decisions are represented in the planner's state, the planner can create plans with a higher likelihood of feasibility. As an example, when an airport is known to be busy, perhaps initial planning of a goal with a tight deadline should advoid the airport. Correctly representing scheduling decisions is complex, because the temporal aspect of the scheduling decision also must be represented. The airport will not always be busy. Thus representing scheduling information when re-planning from a failed plan always seems reasonable, it is unclear how to usefully represent the scheduling information for new plans.
Besides specific data about resource utilization, It might also be useful to represent in the planner general resource information so that the planner could decide what would be a reasonable abstraction level to use for planning, or even guide how to interact with the scheduler. For example, in most cases the planner shouldn't have to pick which airport to use in transporting some goods. However, if airports in one part of the country are very busy then perhaps the planner should investigate shipping the same type of goods from a different part of the country where the airports are more free. Thus because of scheduling information, the planner considers location information that it didn't consider before.
The details of representing scheduling information in the planner and how it might affect the planning will be given in the two sections below. However, at this time it is worthwhile to point out that if this information is only used when re-planning on scheduling failures, the representation is much simpler because the persistence of the scheduling information is just the re-planning time.