O-Plan Current and Future

Introduction

The aim of the article is to provide an overview of the current state of the O-Plan project and to describe the future direction of the next phase of the system's development. The next phase of the O-Plan project started on July 15th 1995 and is funded under Phase III of the DARPA/Rome Laboratory Planning and Scheduling Initiative (ARPI). The ARPI is a major US Initiative aimed at providing the next generation intelligent planning and scheduling tools to the US military and has been funded to a level of 65 million dollars over its three phases.

The overall aim of the O-Plan project is to demonstrate how a planner, situated in a task assignment and plan execution (command and control) environment, and using extensive domain knowledge, can allow for flexible, distributed, collaborative, and mixed-initiative planning. The research is seeking to verify this total systems approach by studying a simplified three-level model with separable task assignment, plan generation and plan execution agents.

Current O-Plan

O-Plan is a command, planning and control architecture with an open modular structure intended to allow experimentation on, or replacement of, various components. The research funded under ARPI Phase II sought to determine which functions are generally required in a number of application areas and across a number of different command, planning, scheduling and control systems.

The outcome of the research has achieved a clearer understanding of the components necessary in a flexible planning system, and has shown how such components can be combined in an open systems integration architecture. The work has determined improved ways in which the knowledge available from modelling an application domain can be used effectively to restrict search in a planner. An improved characterisation of a plan as a set of constraints on activity opens up many possibilities for richer distributed, cooperative and mixed-initiative planning systems in the future. The project has created a prototype implementation which has been demonstrated on a class of realistic applications.

The Future of O-Plan

The research and development in the new O-Plan project addresses the development and refinement of a number of Courses of Action (COAs). The COAs are based on different assumptions concerning a developing crisis, the assets to be committed to the mission and the level of response which is being considered. The overall O-Plan research is concerned with three levels:

• task assignment and mission description
• planning
• plan execution support

The new project concentrates on the first two levels and the communication between them. The approach assumes high levels of user interaction within the COA development process, but allows for system initiative where appropriate. Presented below is a simplied view of the interaction between two people who have different roles in the planning process. These are:

• Task Assigner Role: A person supporting the Commander in Chief (CinC) who is concerned with developing briefings on alternative COAs and ensuring the recommendations on appropriate responses to an emerging crisis are available for selection and subsequent execution.

• Planner Role: A person developing and refining one or more COAs to a given level of detail within a brief specified by the Task Assigner.

Each person is supported notionally by a separate computer system and interaction between the two is in terms of ``planning task flow''. The planning task flow between the task assigner and the planner is demonstrated in the following partial scenario.

• task assigner asks for the development of 3 to 4 COAs to a given level of detail for different levels of response to the emerging crisis, with initial objectives identified and with a safe and low estimate of available resources

• planner generates initial COAs using the outline plan, objectives and constraints available in the plan library for each level of response indicated.

• task assigner evaluates the plans against various elements of evaluation and in particular identifies the potential date at which various mission critical milestones will occur.

• planner improves the evaluation criteria to be used when selecting between alternatives considered in its search for the COAs.

• task assigner for certain COAs adds in specific plan milestones required and may add new objectives.

• planner working with system support refines the plan to account for the additional requirements.

• task assigner receives emerging intelligence may provide a clearer picture of the situation, e.g. better estimates of the date of critical milestones, objectives to be achieved, availability and location of available assets, etc. This information may invalidate one or more of the emerging COAs or require them to be extended.

• planner develops new COAs as necessary using tasking information provided or refines existing COAs to cope with the added requirements.

• task assigner seeks more detailed plan information from briefings and presentations.

• planner a chosen COA to a greater level of detail for nominated phases to establish the feasibility of the plan.

• task assigner asks for a number of specific changes to the chosen COA to deal with issues raised during the briefing.

• planner through user/system dialogue develops a number of alternative detailed plans. Some may meet the new objectives but introduce negative factors. Others may not meet all objectives but stay within other tasking requirements.

• task assigner through briefings (supported by the plan representations and analysis information) selects a particular COA.

• planner revises the chosen COA to reflect updated asset location information and crisis intelligence.

This outline scenario traces the interaction between the roles task of the assigner and the planner in a military planning scenario. The user and system components involved will be working together to define, populate and evaluate an alternative COA comparison matrix which relates important and relevant domain evaluation criteria.

In the current O-Plan system the research has been anchored around the simple dialogue between the task assigner, planner and execution support system and the orders issued and authorities provided within the military command, planning and control environment. The management of dialogue between the task assigner and the planner and their integration into a mixed initiative plan generation and refinement environment will be the one of the main areas of research on the project.

A shared representation for task and plans is vital to the success of integrated planning, scheduling and resource management tools. It can enable shared plan use across components of the ARPI and can influence future planning aids for military and business sectors. Research on the new O-Plan project will play a large part in progressing the development of a shared planning ontology and plan representations suited to dual use across many types of organisation involved in planning.

Brian Drabble: b.drabble@ed.ac.uk
Planning and Scheduling Group