Abstract
Plan generation is essentially a process of reasoning about the future. In complex application conditions, humans need computerized support by planning assistant systems that manage the complexity of data and the dynamics of targeted processes. In practice, information on real processes and human behavior is typically incomplete, maybe vague, and sometimes incorrect. Consequently, realistic planning is more inductive than deductive and the plans generated by an assistant system are hypothetical. In monotonic approaches to plan generation, intermediate plans are increasingly better fitting the application needs over time. Monotonicity is some logical property of the planning process. It depends on a variety of conditions which have to be balanced. Solutions for monotonic plan generation are elaborated in detail.