c7401864

Guide
- Agents and actions
- Authoring a behaviour tree
- Behaviour trees
- The blackboard
- The AI debugger
- Environment queries
- GOAP
- Perception
- Sensors
- Utility
- Nodes over the world
- Move sets
- Pose constraints
- The variation harness
- Getting content into a running game
- Fuzzing a format that has a grammar
- Symbols
- Component defaults
- Components
- Entity queries
- Booleans and handoff
- The editing pipeline
- Editor scripts
- Element selection
- Face materials
- Foliage mode
- The frame panel
- Icons
- The editor shell
- Inspectors in markup
- Editable meshes in a scene
- Editor modes
- Building to a number
- Scene menus
- The shape tool
- Snapping
- Sub-object picking
- Sculpt and paint mode
- Water mode
- Writing an editor plugin
- Block-out shapes
- Booting an application
- Floating things on water
- Character movement
- Editable meshes
- Growing a forest
- Foliage instances
- Grass
- Mesh booleans
- Mesh operations
- Mesh surfaces
- Networked players
- Players and possession
- Splines
- Terrain brushes
- The terrain heightfield
- Painting a terrain
- Sculpting a heightfield
- Where the water surface is
- Saving and restoring a world
- Leashing and spawn tables
- Attributes and the modifier algebra
- Chat channels, routing and moderation
- Collections, achievements and transmog
- Abilities and damage
- Definitions and the content walk
- Dynamic events and world bosses
- Currencies, vendors, trade, mail and the auction
- Effects
- Gameplay events
- Player housing
- Dungeons, raids and lockouts
- Interactables, gathering and crafting
- Inventory and the container algebra
- Items
- Authoring a loot table
- Loot tables
- Gameplay modules
- Mounts and vehicles
- Levels, talents and reputation
- Arenas, battlegrounds and objectives
- The quest and event-chain editor
- Quests, stages and objectives
- Gameplay randomness
- Requirements
- Weapons and hit claims
- Parties, guilds, friends and presence
- Gameplay tags
- Maps, discovery and travel
- Admission, health and the control plane
- Durable state and the ledger
- The gameplay bridge
- Matchmaking, queues and ratings
- Placing players
- Placing realms
- Shards, keys and specs
- The cluster
- The live verbs, at three in the morning
- The service plane
- Transfer tickets
- Transferring players
- Upgrading a fleet
- Writing a realm
- Choosing a frame
- Drawing foliage
- The ambient split, and the nodes that fill it
- Drawing grass
- Impostors
- Culling and streaming instances
- Turning on dynamic global illumination
- The look profile
- Meshes and materials, type by type
- Drawing particles
- Lighting a scene in lux and lumens
- Making a room look different
- The post-processing node kinds
- Tracing rays against the scene's own triangles
- A pass that reads the frame so far
- Render quality presets
- Making everything cast a shadow
- The Standard Frame
- Drawing a terrain
- Streaming texture mip tails
- Timing the frame on the GPU
- Fog a shadow can fall through
By kind
- Components48
- Systems55
- Controls169
- Shaders63
- Graph nodes64
- Importers28
- Annotations45
- Diagnostics8
- Log events69
All namespaces315
Releases1

IAgentGovernor

interface

Core/Vixen.Ai/Agents/IAgentGovernor.cs:100

csharp

public interface IAgentGovernor

Decides which agents may think this tick.

Read the guide page for this →

Remarks

Behaviour-tree steps are cheap and utility scoring is bounded, but neither is free, and a thousand agents doing both every frame is a frame nobody gets back. The governor is the one place that decides who is worth it, and it sits above all three planners because doc 37 § D2 made the agent one shape.

⚠ Plan must be a pure function of its arguments. An amortised scheduler is time-dependent by construction, which is a real hole in determinism — so the hole is bounded by making the schedule reproducible: given the same tick and the same population, every machine picks the same agents. Not arrival order, not a queue, not a priority sort on a float. docs/plan/37 § D18 states this rather than leaving it to be discovered as a desync six months in, and AgentGovernorTests asserts it.

Methods (1)

AgentSchedule Plan(long tick, int population)
Works out this tick's window.

Used by (6)

AiSchedulingTestsVixen.Ai.Tests
AiSystemVixen.Ai
AiSystemTestsVixen.Ai.Tests
RoundRobinGovernorVixen.Ai
SeamTestsVixen.Ai.Nodes.Tests
UnboundedGovernorVixen.Ai