Charles L. Ortiz

A Survey of Research in Distributed, Continual Planning

Complex, real-world domains require rethinking traditional approaches to AI planning. Planning and executing the resulting plans in a dynamic environment implies a continual approach in which planning and execution are interleaved, uncertainty in the current and projected world state is recognized and handled appropriately, and replanning can be performed when the situation changes or planned actions fail. Furthermore, complex planning and execution problems may require multiple computational agents and human planners to collaborate on a solution. In this article, we describe a new paradigm for planning in complex, dynamic environments, which we term distributed, continual planning (DCP). We argue that developing DCP systems will be necessary for planning applications to be successful in these environments. We give a historical overview of research leading to the current state of the art in DCP and describe research in distributed and continual planning.

Distributed Sensor Networks

This report describes the work performed on the DARPA Distributed Sensor Networks Program at Lincoln Laboratory during the period 1 April through 30 September 1986.

Distributed multirobot exploration, mapping, and task allocation

We present an integrated approach to multirobot exploration, mapping and searching suitable for large teams of robots operating in unknown areas lacking an existing supporting communications infrastructure. We present a set of algorithms that have been both implemented and experimentally verified on teams—of what we refer to as Centibots—consisting of as many as 100 robots. The results that we present involve search tasks that can be divided into a mapping stage in which robots must jointly explore a large unknown area with the goal of generating a consistent map from the fragment, a search stage in which robots are deployed within the environment in order to systematically search for an object of interest, and a protection phase in which robots are distributed to track any intruders in the search area. During the first stage, the robots actively seek to verify their relative locations in order to ensure consistency when combining data into shared maps; they must also coordinate their exploration strategies so as to maximize the efficiency of exploration. In the second and third stages, robots allocate search tasks among themselves; since tasks are not defined a priori, the robots first produce a topological graph of the area of interest and then generate a set of tasks that reflect spatial and communication constraints. Our system was evaluated under extremely realistic real-world conditions. An outside evaluation team found the system to be highly efficient and robust.

A commonsense language for reasoning about causation and rational action

Abstract Commonsense causal discourse requires a language with which to express varying degrees of causal connectedness. This paper presents a commonsense language for reasoning about action and causation whose semantics is expressed by way of counterfactuals. Causal relations are analyzed along several dimensions including notions of resource consumption, degree of responsibility, instrumentality, and degree of causal contribution. Grounding the semantics in a level of counterfactual reasoning is shown to play an important role in constraining the set of allowable event descriptions instantiating reports expressed by any of the relations in the language. These ideas are also applied to a causal analysis of rational action: by adopting an explanatory stance, one can characterize action through descriptions that refer to causal connections between mental states and actions. Such a causal analysis resolves some well-known difficulties in correctly ascribing agency and intentionality. Finally, an implementation is described—used to motivate and refine the theory—in which queries involving causal relations between the activities of agents engaged in purposeful behavior within a microworld can be posed.

Interpreting Information Requests in Context A Collaborative Web Interface for Distance Learning

We describe the use of theories of agent collaboration and human dialogue processing in providing a principled basis for the design of web interfaces to multimedia information stores. The DIAL system, an implementation in the domain of information support for distance learning by students in an introductory programming class, is used to illustrate the efficacy of this approach. DIAL builds a representation of context that is based on the collaborative plans of the system and its user and uses this contextual information to reduce the communication burden. Context is represented by a structure of intentions that a user is attempting to satisfy. This structure is modified as tasks are completed or task descriptions are refined. DIAL interprets information requests relative to the prevailing context as it is represented by this structure. As a result, requests may be expressed more economically; contextual information is added by the system. Furthermore, DIAL uses information about the intentional context to respond and act collaboratively, rather than in the master-slave style typical of most current human-computer interfaces. DIAL and the access method it supports provide a unique support tool for distance learning environments as well as a demonstration of a general way in which agent models can be used to improve human-computer communication.

Dynamic intention structures I: a theory of intention representation

This article introduces a new theory of intention representation which is based on a structure called a Dynamic Intention Structure (DIS). The theory of DISs was motivated by the problem of how to properly represent incompletely specified intentions and their evolution. Since the plans and intentions of collaborating agents are most often elaborated incrementally and jointly, elaboration processes naturally involve agreements among agents on the identity of appropriate agents, objects and properties that figure into their joint plans. The paper builds on ideas from dynamic logic to present a solution to the representation and evolution of agent intentions involving reference to incompletely specified and, possibly, mutually dependent intentions, as well as the objects referenced within those intentions. It provides a first order semantics for the resulting logic. A companion paper extends further the logical form of DISs and explores the problem of logical consequence and intention revision.

Distributed Sensor Networks: Introduction to a Multiagent Perspective

As computer networks (and computational grids) become increasingly complex, the problem of allocating resources within such networks, in a distributed fashion, will become more and more of a design and implementation concern. This is especially true where the allocation involves distributed collections of resources rather than just a single resource, where there are alternative patterns of resources with different levels of utility that can satisfy the desired allocation, and where this allocation process must be done in soft real-time. This book is the first of its kind to examine solutions to this problem using ideas taken from the field of multiagent systems. The field of multiagent systems has itself seen an exponential growth in the past decade, and has developed a variety of techniques for distributed resource allocation.

Structured negotiation

Structured negotiation is proposed as a new method through which collaborating agents can seek consensus on the apportionment of tasks and resources. The approach draws on research in collaborative planning and human dialog understanding: agent interactions are organized in a manner that reflects the structure of a shared plan. Negotiations are incremental and interleaved with the shared planning process while communications supporting negotiations are made efficient by drawing on knowledge of a prevailing context. Agent proposals to team members are annotated with causal information that compactly expresses relationships between new proposals and the current context. Normative guidelines for proposal generation further restrict communications of ancillary information to only those fragments that represent departures from the norm. Finally, a set of interpretation rules allows agents to infer information not explicitly communicated.

Dynamic Resource-bounded Negotiation in Non-additive Domains

The problem of group decision making in a non-strategic environment is presented and analyzed. The main focus is on the decision problem of task assignment in situations in which tasks interact and information relevant to the assignment problem is distributed and is contained locally in the agents in the group. Practically, it may be impossible to communicate all relevant distributed information to a single central decision maker due to communication costs, the size of the set of information, or other limitations. Instead, we provide a method to coordinate the sharing of a limited amount of information while making satisfactory (though possibly suboptimal) task assignments via a center-based algorithm called Mediation. Mediation implements an iterative and interactive hill-climbing search in a subset of the solution space by making successive proposals and sending those proposals to the group. Each proposal provides a context on which group members base their responses which provide the mediator with information to find a satisfactory outcome to the assignment problem. The properties of Mediation are compared with other approaches including parallel and combinatorial auctions. The theory and analysis is illustrated with examples from the domain of multi-sensor intruder tracking. Dynamic mediation extends the algorithm to environments in which problem features change during the decision-making process and in which agents augment the information that they provide using the language of rich bids. Experiments are used to validate the usefulness of mediation in key problem domains, including multi-sensor tracking. Finally, an architecture for agents, who need not be stationary, is described whereby agents can monitor task progress at execution time and then modify existing resource allocations based on the evolving situation.

Hierarchical information combination in large-scale multiagent resource management

In this paper, we describe the Distributed Dispatcher Manager (DDM), a system for managing resource in very large-scale task and resource domains. In DDM, resources are modeled as cooperative mobile teams of agents and objects or tasks are assumed to be distributed over a virtual space. Each agent has direct access to only local and partial information about its immediate surroundings. DDM organizes teams hierarchically and addresses two important issues that are prerequisites for success in such domains: (i) how agents can extend local, partial information to arrive at a better local assessment of the situation and (ii) how the local assessments from teams of many agents can be integrated to form a global assessment of the situation. We conducted a large number of experiments in simulation and demonstrated the advantages of the DDM over other architectures in terms of accuracy and reduced inter-agent communication.