Representational Requirements Functional Requirements Back to CPF

Representational Requirements

• Activities
  • Abstract activity
  • Activity arguments
  • Activity decompositions (more than one possible)
  • Activity orderings
  • Activity templates
  • Activity variables (not arguments)
  • Activities with context-dependent effects
  • Activities with indirect effects (i.e. ramifications)
  • Alternative activities
  • Complex activities / knowledge producing activities (i.e. activities including program constructs)
  • Complex group of activities
  • Complex sequences
  • Concurrent activities
  • Condition types: supervised, unsupervised, achieve, only-for-query etc
  • Hierarchical decomposition of activities
  • Non-deterministic activities
  • Parallel activities
  • Postconditions (Add and delete lists)
  • Preconditions
  • Primitive activity
  • Relationship between activities and goals
  • Relationship between activities and resources
  • Relationship between activities and agents
  • Relationship between activities and time/space
  • Role of activities (e.g. sub-activity)
  • Simple activity representation
  • Simple sequence
  • Simple grouping of activities
• Agents
  • Agent assumptions
  • Agent beliefs
  • Agent capabilities
  • Agent desires
  • Agent intentions
  • Agent model
  • Agent performing an activity
  • Agent purposes (i.e. purposes held by an agent)
  • Agent roles
  • Agent's preferred alternatives for plans and plan elements
• Control Structures/Execution/Simulation
  • Conditionals
  • Detailed simulation information
  • Executability flag on a plan
  • Execution errors
  • For each, iterators
  • Loops, cycles
  • Monitors (monitor the world)
  • Protection intervals for conditions to be maintained
  • Recovery procedures that will allow recovery from failure states
  • Reflective reasoning support (system goals and activity)
  • Represent event-driven behaviour (how to respond to specific events)
  • Simple simulation information
• Domain Knowledge
  • Always true facts
  • Applicability of a procedure or operator to accomplish a goal
  • Conditions which must be met before a procedure or operator can be used
  • Deduced consequences of changes in the world
  • Domain assumptions
  • Domain in which certain activities take place
  • Domain model (i.e. specifying possible behaviours)
  • Global constraints
  • Information required for deductive reasoning
  • Operators, cases, ACTs
  • Relation to those things not in the plan model (ie the environment)
  • Specification of activities relative to a domain
  • World conditions (precond/effect)
• Evaluations
  • Cost data
  • Evaluation criteria
  • Level of effort
  • Meta-data that characterises conditions for completeness and state of completeness (e.g. as in blackboard systems)
  • Resource utilisation
• General Structures
  • Assertions/retractions
  • Frames
  • Fuzzy rules, fuzzy sets
  • General constraints - hard/softness of a constraint, must be true, should be true, cannot be true
  • Measurements/amounts: size, weight, time, cost
  • Nodes and arcs
  • Numbers
  • Queries
  • Queues, stacks ,lists
  • Relations
  • Sentences (boolean)
  • Sets
  • Symbols
  • Terms and expressions
• Goals, Requirements, Objectives, Mission, Task (GROMT)
  • Branching to labelled goals
  • Conditional goals
  • Constraints on achievement of the goal/task, etc
  • Goals having different modalities
  • Goals of achievement
  • Goal of maintenance
  • Goals of testing
  • Goal ordering
  • Phantom goal
  • Satisfied goal
  • Simple Goal, Objective, Requirement, Mission, Task Representations
  • Task time limits
  • Unsatisfied goal
• Organisational
  • Authority (conditions, permissions, grant/revoke)
  • Deadlines
  • Demand (i.e. an input request for one or more products (order))
  • Filters for simplicity (e.g. hide certain detail, etc.)
  • Knowledge of current systems
  • Knowledge of products
  • Organisational model (relationships, hierarchy, etc.)
  • Organisational procedures and policies
  • Manual vs automated activities
  • Milestones
  • Model of the organisation's planning process and planning capabilities
  • Plan constraint owner
  • Published and private planning views
  • Who, What, Where, When, Why as well as How
• Plans/Schedule
  • Abstraction levels in the domain model
  • Ad hoc notes/annotations optionally associated with any component of the plan
  • And-split, and-join
  • Document references (Illustrations/drawings/references, etc.)
  • Dummy nodes (simplify ordering constraint specification, user communication)
  • Extensible (eg Partial shared views (PSV))
  • Issues, work list of things still to address (flaws, unsatisfied conditions, unresolved interactions, abstract activities that still require expansion
  • Or-split, or-join
  • Recursive plan structures
  • Schedule for planned tasks
  • Structural activity decomposition level (different than abst level)
  • Sub-plans
• Rationale
  • Criteria for selecting alternatives (activities, orderings, objects, etc.)
  • Dependencies - between the preconditions and effects of activities involved in the plan, resource provision for other activities, etc.
  • Explanation information
  • Explicit list of option alternatives (activities, orderings, objects, etc.)
  • Explicit list of issues considered
  • Explicit representation of causal structure (teleology) of the plan
  • Historical trace of plan generation
  • Plan choices/decisions (which alternative was chosen)
  • Rationale that supports backtracking (linking plan elements to decisions)
  • What the aims of the plan are/were
• Resources/Objects
  • Actual resources available for a specific task
  • Complex resource characteristics
  • Eligible resources
  • Implicit/explicit resource association
  • Model of general object hierarchies (specialise/generalise) (e.g. "business objects" for military planners)
  • Model that includes the products of planning
  • Object constraints
  • Overcommitments of resources
  • Relationship between resource and goal
  • Resource capability/characteristics
  • Resource categorisation and grouping
  • Resource conditions
  • Resource cost (cost to use it, etc)
  • Resource location
  • Resource requirements and usage for an activity
  • Resource requirements and usage for a task
  • Resource types (consumable, reusable sharable, etc.)
  • Resource/action combined descriptions
  • Resource/Object roles
  • Simple resource representation
• States
  • Changes in world state
  • Individual world state
  • Numerous hypothetical states
  • Simple grouping of states
  • Simple state representation
  • Single world model corresponding to actual perceived state
  • State existence constraints
  • World histories
• Time/Space
  • Activity duration
  • Allen relations
  • Begin/end of an activity
  • Continuous change
  • Duration bounds (minimum duration, max duration)
  • Estimated times
  • Metric, calendar
  • Multiple durations
  • Narrative time
  • Persistence
  • Relative time
  • Spatial point
  • Time (explicitly)
  • Time constraint violations
  • Time feasibility window for activities in plan and resource usage
  • Time interval
  • Time points
  • Temporal constraints
• Uncertainty/Ambiguity
  • Ambiguity
  • Confidence levels
  • Don't know, Don't care (ie ignorance)
  • Imprecision
  • Minimum/maximum information for time windows, resource levels
  • Range of possible outcomes
  • Probablilistic uncertainty
  • Uncertainty
  • Variability/tolerance

Functional Requirements

• Communicate Plan
  • Appreciate plan's strengths and shortcomings
  • Associate meaningful names for the plan levels or phases
  • Communicate plans
  • Communicate to the human what the computer has done
  • Convey ancestry or class of an activity
  • Extend a simple representation to a more complex one
  • Get briefed on plan (to get up to speed quickly)
  • Get oriented to one's task or responsibility in the planning process and/or plan
  • Identifying sensitivities and key dependencies
  • Provide multiple views: activity/sub-activity breakdown structures, PERT charts, Gantt charts, workflow wavefront and todo lists, role activity diagrams and state charts, justifications
  • Provide planning API
  • Provide specialist-specific views on part of the plan they most care about
  • Restrict brief to relevant subset
  • View explicit planning horizon (e.g. let me see the plan for next 24 hrs)
  • Store all the things specific to my specialised type of planning but be able (with some loss) to translate to a representation others can understand
  • Design which assumes object orientation but not a particular language
  • Support human-human, machine-machine and human-machine communication
  • Import/Export parts of a plan
  • Extraneous facets should be automatically hidden
  • Support canonical export formats
• Domain Building
  • Acquire plan domain information
  • Constrain what can be expressed
• Organisational Support
  • Deadline management
  • Delegate part of a plan
  • Determine whether an objective has been met by what see happening
  • Insert/attach milestones
  • Interface with a number of tools/applications
  • Linking plan/task to original requirements or context, circumstances
  • Mark activities as critical
  • Partition up the work for others to work on
  • Passing along requirements to other applications
  • Support workflow management
  • What-if analysis
• Plan Editing/Browsing
  • Add or delete plan elements
  • Annotate hypothetical changes
  • Ask questions for details about the plan
  • Back-track to source of each element in the plan
  • Browse plan states (browse temporal or logical states)
  • Browse a plan
  • Change parameter values
  • Examine the expected effects of current plan parts
  • Examine the rationale of current plan parts
  • Fill in values for parameters
  • Find interesting parts of the plan and look at them simultaneously
  • Find specific information (e.g. keyword search, etc.) or relevant, applicable parts of plan
  • Find out if a decision relevant to a specific perspective has been made
  • Hypothetically make changes to the plan
  • Introspection over the plan
  • Justify additions to the plan to negotiate with others over conflicts
  • Make single step modifications
  • Particularise some generic values
  • Propose/accept plan changes and incorporate them into current plan
  • Provide access to past decision rationales (generate explanation)
  • Provide balance between hand editing of dependencies (as ACPT supports) and computer-provided dependencies, as SIPE and other generative planners provide
  • Provide standard terminology/indexing for information and decisions
  • Support explicitly different levels of abstraction
  • Track changes made to the plan
  • View key dynamics and key outcomes
  • Specialise a design for special planning needs
  • Specialise parts of or throw away parts without affecting parts one cares about
  • Easy and simple to represent easy and simple things
  • Represent information at every stage
  • Able to start by representing just name of a plan
  • Able to name actions without specifying their order but add in ordering later when I have that information
• Plan Execution
  • Exception handling and recovery
  • Exchange info between activities
  • Monitor execution progress
  • Monitor resource usage
  • Record answer from environment via sensor (test result, value, etc)
  • Reflective reasoning
  • Simulate/Animate the plan
  • Synchronise multiple, parallel activity sequences activity
  • Record the state of execution of a plan even though there may be uncertainty about the state
• Plan/Schedule Generation
  • Assign a constraint over an interval
  • Associate resources with activities
  • Associate time intervals or time points with activities
  • Deliberate planning
  • Find an activity to satisfy a goal
  • Find a previous plan that seems applicable to current task
  • Formal reasoning about plan elements
  • Make a goal node a phantom node
  • Merge plans
  • Modify plan (causal-link, MTC-based, backtrack-modify)
  • Non-progressive refinement (partition the search space)
  • Order goals
  • Order sequential/parallel activities
  • Perform math and logical operations
  • Predicting the results of taking certain actions
  • Progressive refinement (add new actions)
  • Propagate constraints
  • Propose expansion (smart, goal achievement, decomposition)
  • Propose refinement
  • Reactive planning
  • Reason about temporal constraints
  • Replace an abstract with a decomposition of sub-activities
  • Respond to new goal and information and take appropriate steps
  • Reuse a plan
  • Reuse components of a plan
  • Roll-back, roll-forward plan modifications
  • Schedule planned tasks
  • See planning progress
  • Select goal (linear, random, smart)
  • Specialise concepts defined in an abstract model or base representation
  • Specify/allocate resources
  • Stub off area to cover incomplete aspects
  • Support uncertainty
  • Transaction-based modifications
  • Trim down representation for simpler cases
• Plan Evaluation/Critique
  • Assess plan fitness
  • Associate evaluation criteria with objectives
  • Associate evaluation criteria with plan
  • Compare plan alternatives
  • Critique interactions (causal-link based, MTC based)
  • Critique plan (consistency)
  • Ensure that intentions were not compromised
  • Evaluate hypothetical plans
  • Evaluate which plan constraints hold in specific world states
  • Perform sensitivity analysis on robustness and risk factors
  • Test for unachieved goals (current state, agenda-based, MTC-based)
  • Validate entire plan
• Planning System Synthesis
  • Configure an executable planning system from specification of an abstract domain model
• Task assignment
  • Associate a constraint with a time point
  • Associate various meaningful names with different task assignments (options), levels, phases
  • Intend, desire, enforce, synthesise a constraint
  • Specify temporal constraints

[Back to CPF]

Date Last Updated: 1 June 1999