Complex modern-day artifacts are designed cooperatively by groups of experts, each with their own areas of expertise. The interaction of such experts inevitably involves conflict. This paper presents an implemented computational model, based on studies of human cooperative design, for supporting the resolution of such conflicts. This model is based centrally on the insights that general conflict resolution expertise exists separately from domain-level design expertise, and that this expertise can be instantiated in the context of particular conflicts into specific advice for resolving those conflicts. Conflict resolution expertise consists of a taxonomy of design conflict classes in addition to associated general advice suitable for resolving conflicts in these classes. The abstract nature of conflict resolution expertise makes it applicable to a wide variety of design domains. This paper describes this conflict resolution model and provides examples of its operation from an implemente...

Abstract not found

In most practical environments, scheduling is an ongoing reactive process where evolving and changing circumstances continually force reconsideration and revision of pre-established plans. Scheduling research has traditionally ignored this "process view" of the problem, focusing instead on optimization of performance under idealized assumptions of environmental stability and solution executability. In this paper, we present work aimed at the development of reactive scheduling systems, which approach scheduling as a problem of maintaining a prescriptive solution over time, and emphasize objectives (e.g., solution continuity, system responsiveness) which relate directly to effective development and use of schedules in dynamic environments. We describe OPIS, a scheduling system designed to incrementally revise schedules in response to changes to solution constraints. OPIS implements a constraint-directed approach to reactive scheduling. Constraint analysis is used to prioritize outstandin...

We have used KIDS (Kestrel Interactive Development System) to derive extremely fast and accurate transportation schedulers from formal specifications. As test data we use strategic transportation plans which are generated by U.S. government planners. In one such problem, the derived scheduler was able to schedule 15,460 individual movement requirements in 71 cpu seconds. The computed schedules use relatively few resources and satisfy all specified constraints. The speed of this scheduler derives from the synthesis of strong problem-specific constraint checking and constraint propagation code. 1 Introduction This paper describes our exploration of the transformational development of transportation schedulers. Our approach involves several stages. The first step is to develop a formal model of the transportation scheduling domain, called a domain theory. Second, the constraints, objectives, and preferences of a particular scheduling problem are stated within a domain theory as a problem...

Abstract not found

Abstract not found

Planning is one of the most important and oldest fields of AI. However, there is no consensus on how to compare and classify planning systems and methods. Neither the traditional view of planning as search nor the formalization efforts have been able to provide a basis for a classification scheme. This article explores the idea that a perspective based on Newell's knowledge level can be useful for this task. We present a knowledge-level analysis of classical planning systems in terms of models of the problem-solving methods they used. Rather than reengineering these systems in detail, however, our goal is to show how this type of analysis can help define which roles knowledge may play in planning tasks, and how these roles can be used to compare planning methods in terms of (i) which types of knowledge are used, (ii) how they are structured in what we call domain models. As a tool to analyze and represent planning methods we use the KADS methodology. 1 Introduction ...

The design of complex artifacts has increasingly become a cooperative process, with the detection and resolution of conflicts between design agents playing a central role. Effective tools for supporting the conflict management process, however, are still lacking. This paper describes a system called DCSS (the Design Collaboration Support System) developed to meet this challenge in design teams with both human and machine-based agents. Every design agent is provided with an "assistant" that provides domain-independent conflict detection, classification and resolution expertise. The design agents provide the domainspecific expertise needed to instantiate this general expertise, including the rationale for their actions, as a part of their design activities. DCSS has been used successfully to support the cooperative design of Local Area Networks by human and machine-based designers. This paper includes a description of DCSS's underlying model and implementation, examples of its operation...

In a joint project, BBN and Kestrel Institute have developed a prototype of a mixed-initiative scheduling system called ITAS (In-Theater Airlift Scheduler) for the U.S. Air Force, Pacific Command. The system was built in large part using the KIDS (Kestrel Interactive Development System) program synthesis tool. In previous work for the ARPA/Rome Laboratory Planning Initiative (ARPI), Kestrel has used their program transformation technology to derive extremely fast and accurate transportation schedulers from formal specifications, as much as several orders of magnitude faster than currently deployed systems. The development process can produce highly efficient code along with a proof of the code's correctness. This paper describes the current prototype ITAS system and its scheduling algorithm, as a concrete example of a generated scheduling working on a real problem. We outline the generated search algorithm in order to promote and facilitate comparison with other constraint-based schedu...

Formal methods are usually conceived as a way to obtain veri ably correct software, so many researchers have focused on applications requiring error-free code, such as safety-critical subsystems. There may be other paths to the ultimate success of formal methods. We argue that mechanized synthesis tools can have an impact in the production of high-performance algorithms. This thesis is supported by our work on the synthesis of transportation scheduling algorithms. 1