Home > Tasks > Reasoning > AnalogicalReasoningTask

AnalogicalReasoningTask

Solve complex problems by mapping structural relationships and insights from diverse source domains (e.g., biological systems, urban planning, music).

Reasoning GPT-4 Preferred Non-Destructive

⚙️ AnalogicalReasoningConfig.json

{
  "source_domain": "Biological Systems",
  "target_problem": "Scaling a distributed microservice architecture",
  "num_analogies": 3,
  "validate_mappings": true,
  "input_files": ["docs/architecture_spec.md"],
  "related_files": ["src/main/kotlin/com/system/Core.kt"]
}

→

👁️ Session UI: Synthesis & Recommendations

Synthesis: Mycelial Network Analogy

Key Insight: Decentralized resource allocation via chemical signaling mimics service discovery in high-latency environments.

Conceptual Mapping:

                        Fungal Hyphae → Network Routes

                        Nutrient Flow → Data Packets

                        Spore Distribution → Container Deployment

Recommended Approach: Implement a "Gossip Protocol" for state propagation based on the Mycelial signaling model to reduce central orchestrator load.

Workspace Explorer

Browse generated reasoning artifacts, transcripts, and synthesis reports in the task workspace.

TaskExecutionConfig Fields

Field	Type	Description
source_domain *	String	The domain to draw analogies from (e.g., 'biological systems', 'urban planning').
target_problem *	String	The specific problem to solve using analogical reasoning.
num_analogies	Int	Number of analogies to generate and explore. Default: `3`.
validate_mappings	Boolean	Whether to perform a secondary validation pass for structural consistency.
input_files	List<String>	Glob patterns for files providing context for the target problem.
related_files	List<String>	Additional context files to inform the reasoning process.

Token Usage: Medium-High (Depends on num_analogies and context size).

Kotlin Boilerplate

// 1. Define the Execution Configuration
val executionConfig = AnalogicalReasoningTaskExecutionConfigData(
    source_domain = "Biological Systems",
    target_problem = "Optimizing data flow in a mesh network",
    num_analogies = 3,
    validate_mappings = true,
    input_files = listOf("docs/network_spec.md")
)
// 2. Run via UnifiedHarness (Embedded Mode)
harness.runTask(
    taskType = AnalogicalReasoningTask.AnalogicalReasoning,
    typeConfig = TaskTypeConfig(name = "NetworkOptimizer"),
    executionConfig = executionConfig,
    workspace = File("./workspace"),
    autoFix = true
)

Embedded Integration (Headless)

To use this task in a CI/CD pipeline or CLI tool, initialize the UnifiedHarness in serverless mode:

val harness = UnifiedHarness(
    serverless = true,
    openBrowser = false,
    modelInstanceFn = { apiModel -> 
        apiModel.model!!.instance(key = System.getenv("OPENAI_API_KEY")) 
    }
)
harness.start()

Test Case Example

Scenario: Urban Planning for Microservices

Source Domain	Urban Traffic Management
Target Problem	Reducing latency in a distributed event-bus architecture
Expected Insight	Implementing "Roundabouts" (buffer zones) for high-traffic event streams to prevent deadlock.

CLI Invocation

# Example using the Cognotik CLI JAR
java -jar cognotik-cli.jar \
  --task AnalogicalReasoning \
  --param source_domain="Music Theory" \
  --param target_problem="Refactoring a complex inheritance hierarchy" \
  --workspace ./my-project

Prompt Segment

The following instructions are injected into the LLM context to guide the reasoning process:

Prompt Segment

AnalogicalReasoning - Solve problems by finding and applying analogies from different domains
** Specify a source domain to draw analogies from (e.g., biological systems, architecture, music)
** Provide the target problem you want to solve
** Configure the number of analogies to generate (default: 3)
** Optionally enable mapping validation for structural consistency
** The task will:
   - Identify relevant concepts in the source domain
   - Map structural relationships to the target problem
   - Generate insights and potential solutions
   - Validate the coherence of the analogical mappings
   - Synthesize findings across multiple analogies

Synthesis: Mycelial Network Analogy

Workspace Explorer

TaskExecutionConfig Fields

Task Process Lifecycle

Kotlin Boilerplate

Embedded Integration (Headless)

Test Case Example

CLI Invocation

Prompt Segment

Prompt Segment