Intelligence Plane — What To Do, When, and How¶

Purpose¶

The Intelligence Plane turns goals into actions. It decomposes high-level objectives into tasks (Planner), allocates resources and time (Scheduler), and orchestrates execution while handling connectivity and failure (Execution Manager). This is the “brain” of the platform.

Services¶

Service	Responsibility	Detail
Planner	Strategic task decomposition and goal reasoning	Multi-resource, order-level, cloud-side planning
Scheduler	Resource allocation and time optimization	One-time, recurring, and conditional scheduling
Execution Manager	Orchestration and monitoring of task execution	Execution contract lifecycle, edge reconciliation

Pipeline¶

  Goal/Order                                                   Device
      │                                                          ▲
      ▼                                                          │
 ┌──────────┐    tasks    ┌───────────┐  assignments  ┌──────────────────┐
 │ Planner  │ ──────────► │ Scheduler │ ────────────► │Execution Manager │
 └──────────┘             └───────────┘               └──────────────────┘
      │                        │                             │
      │ consults               │ consults                    │ pushes
      ▼                        ▼                             ▼
 Policy Service           Policy Service              Execution Contract
 Marie (spatial)          DeviceAdmin                 ThingIO (monitor)

Planner receives a goal and decomposes it into an ordered set of tasks.
Scheduler assigns tasks to resources based on availability, proximity, and policy constraints.
Execution Manager pushes execution contracts to devices, monitors progress via ThingIO (telemetry data), and handles the live/projected/reconciled state machine.

Two-Tier Planning¶

Tier	Location	Scope	Nature
Platform Planner	Cloud (FlexGalaxy.AI)	Global, multi-resource	Strategic — “which resources handle which tasks”
Edge Planner	Local (robot/PDA)	Single-agent, local	Tactical, reactive — “obstacle ahead, reroute”

These are fundamentally different planners solving different problems. They do not share logic. The Execution Contract bridges them — the platform tells the edge what to achieve, the edge decides moment-to-moment how.

Data Ownership¶

Data	Owner	Consumers
Task decompositions, plans	Planner	Scheduler
Schedules, resource assignments	Scheduler	Execution Manager
Execution contracts, state machine	Execution Manager	Devices, Applications