DecisionIsolateMessage class

Message payload for communication between main isolate and worker isolate.

This class encapsulates all information needed to execute a decision-making task in a separate isolate and return results to the main isolate.

Isolate Communication Pattern:

In Dart, isolates don't share memory - they communicate by passing messages. This class serves as the standardized message format for all decision-making operations, ensuring type-safe and structured communication.

Message Flow:

Main Isolate                    Worker Isolate
    |                                |
    |---> DecisionIsolateMessage --->|
    |     (algorithm, command, data) |
    |                                |
    |                           [Process]
    |                                |
    |<--- Result via replyPort <-----|
    |                                |

Why Isolates?

Heavy computational tasks (matrix operations, eigenvector calculations) can block the UI thread. By running these in isolates:

• UI remains responsive
• Utilizes multiple CPU cores
• Prevents app freezing during calculations

Architecture:

1. Main isolate creates message with task details and reply port
2. Worker isolate receives message, processes task
3. Result is sent back through the reply port
4. Main isolate receives result and continues execution

Fields:

• algorithm: Specifies which decision-making algorithm to use (AHP or SAW)
• command: The specific operation within that algorithm to execute
• payload: Input data required for the operation (matrices, vectors, etc.)
• replyPort: SendPort for the worker to send results back to the main isolate

Example Usage:

// In main isolate
final receivePort = ReceivePort();

final message = DecisionIsolateMessage(
  algorithm: DecisionAlgorithm.ahp,
  command: AhpProcessingCommand.calculateEigenVectorCriteria,
  payload: {
    'matrix': [
      [1.0, 3.0, 5.0],
      [0.33, 1.0, 2.0],
      [0.2, 0.5, 1.0]
    ]
  },
  replyPort: receivePort.sendPort,
);

// Send to worker isolate
workerSendPort.send(message);

// Wait for result
final result = await receivePort.first;
print('Eigenvector: $result');

Supported Algorithms:

• DecisionAlgorithm.ahp: Analytic Hierarchy Process
• DecisionAlgorithm.saw: Simple Additive Weighting

AHP Commands:

• generateInputPairwiseAlternative: Create comparison templates
• generateResultPairwiseMatrixCriteria: Build criteria matrix
• generateResultPairwiseMatrixAlternative: Build alternative matrix
• calculateEigenVectorCriteria: Compute criteria weights
• calculateEigenVectorAlternative: Compute alternative priorities
• checkConsistencyRatio: Validate comparison consistency
• calculateFinalScore: Compute final decision scores

SAW Commands:

• generateSawMatrix: Create decision matrix
• normalizeMatrix: Normalize values for calculation

Payload Requirements:

The payload map must contain all data required by the specific command. Structure varies by command - refer to individual command documentation for required keys and value types.

Error Handling:

If the worker encounters an error, it sends back a map with 'error' key:

{
  'error': 'Error message here',
  'stack': 'Stack trace here'
}

The main isolate should check for this error format and handle appropriately.

Performance Considerations:

• Message passing has overhead - use isolates for computationally heavy tasks
• Small calculations may be faster on main isolate
• Threshold typically: matrix size > 80 or criteria count > 25
• Serialization cost increases with payload size

Thread Safety:

This message is immutable by design (all fields are final). Once created, it cannot be modified, ensuring thread-safe communication.