Multi-Party Agreement Reactive Decisions

Step ID	Type	Stage	Trigger Type	Specific Trigger	Past Context (Examples)	Available Actions	Selection Function	Next States	History Components	Policy Type	Forward Output	Backward Flow
A1-initiation	DECISION	Initiation	Process Events	agreement_needed	Agreement requirements Own capabilities Resource availability Known participants Market conditions	Initiate agreement Don't initiate	argmax ε-greedy(ε) softmax(τ) system_rule other	Initiate → A2 Don't initiate → Terminal: Waiting	Past initiations Success rates Agreement outcomes Resource utilization	Learned initiation model	Action	Utility Information
A2-evaluation	DECISION	Consensus	Network Events	proposal_received	Proposal terms Current signatures Agreement fit Commitment requirements Other participants	Sign and Potentially Commit Resources Modify terms Reject proposal	argmax ε-greedy(ε) softmax(τ) system_rule other	Sign → Threshold Check → A3 | A2 Modify → A2 (new version) Reject → Terminal: Not Participant	Past evaluations Signature decisions Agreement outcomes Reward realizations	Learned evaluation model	Action	Utility Information
S1-threshold-check	AUTOMATIC	Consensus	Process Events	signatures_updated	Signature count Threshold requirement Time elapsed Participant capabilities	Automatic evaluation Met → consensus Not met → continue Timeout → fail	system_rule	Met → A3 Not met → A2 Timeout → Terminal: Failed	Past thresholds Formation times Participant counts Success rates	System rule	Status	N/A
S2-activation-mode	AUTOMATIC	Activation	Process Events	consensus_reached	Commitment types Agreement terms Participant preferences Resource locks	Automatic evaluation Manual mode Auto-binding mode	system_rule	Manual → A3 Auto-binding → Active Agreement	Agreement types Activation patterns Commitment levels Success correlations	System rule	Mode	N/A
A3-activation	DECISION	Activation	Process Events	consensus_reached	Final participants Agreement terms Resource requirements Commitment status	Activate agreement Abort formation (Skip if auto-binding)	argmax ε-greedy(ε) softmax(τ) system_rule other	Activate → Active Agreement Abort → Terminal: Aborted Skip → Active Agreement (auto)	Past activations Agreement success Resource locks Formation costs	Learned activation model	Action	Utility Information
	Terminal

Decision Summary

A1: Initiation - Agent decides whether to initiate multi-party agreement based on coordination needs, participant availability, and expected benefits versus coordination costs.

A2: Evaluation - Agents evaluate received proposals, deciding to sign (optionally with committed resources), modify terms, or reject based on capability match and participant quality. Can iterate through multiple proposal versions.

A3: Activation - Agent decides whether to activate finalized agreement or abort, considering participant reliability and resource availability. Only occurs if activation mode is manual.

S1: Threshold Check - System automatically monitors signature accumulation, checking if consensus threshold is met, timeout exceeded, or circulation should continue.

S2: Activation Mode - System determines whether manual activation decision is needed or auto-binding applies based on commitment types and agreement complexity.