# Dcsn Optimization Summary DMO

Stores cycle-level (policy-level) performance analytics for decision optimizers (contextual bandits). Each record aggregates across all arms within a single evaluation cycle, providing total pull counts, cumulative regret, best-arm reward rate, and weighted average performance.

## Object Properties

**Object API Name:** `DcsnOptimizationSummary_std__dlm`\
**Category:** Unassigned\
**Availability:** Available in 262 and later versions\
**Primary Key Field:** Id

## Fields

- [AvgModelCalibrationScoreNbr](#avgmodelcalibrationscorenbr)
- [BestOptionRewardRateNbr](#bestoptionrewardratenbr)
- [cdp_sys_record_currency](#cdp_sys_record_currency)
- [CumulativeRegretNumber](#cumulativeregretnumber)
- [DataSourceId](#datasourceid)
- [DataSourceObjectId](#datasourceobjectid)
- [DcsnOptmztnIdentifier](#dcsnoptmztnidentifier)
- [EvaluationCycleIdentifier](#evaluationcycleidentifier)
- [EvaluationDateTm](#evaluationdatetm)
- [Id](#id)
- [InternalOrganizationId](#internalorganizationid)
- [KQ_Id](#kq_id)
- [MetricName](#metricname)
- [OptionCount](#optioncount)
- [TotalCumulativeRewardNumber](#totalcumulativerewardnumber)
- [TotalExplorationRatePct](#totalexplorationratepct)
- [TotalPullCount](#totalpullcount)
- [WeightedAvgRewardRateNbr](#weightedavgrewardratenbr)

### AvgModelCalibrationScoreNbr

- **Field API Name:** `AvgModelCalibrationScoreNbr__c`
- **Data Type:** DOUBLE
- **Description:** Pull-count-weighted average of ModelCalibrationScore across all arms: sum(PullCount times CalibScore) / sum(PullCount); lower values indicate better overall model calibration

### BestOptionRewardRateNbr

- **Field API Name:** `BestOptionRewardRateNbr__c`
- **Data Type:** DOUBLE
- **Description:** Maximum EmpiricalRewardRate across all arms in the cycle; represents the performance ceiling

### cdp_sys_record_currency

- **Field API Name:** `cdp_sys_record_currency__c`
- **Data Type:** TEXT
- **Description:** System-generated metadata field that stores a 3-letter ISO currency code.

### CumulativeRegretNumber

- **Field API Name:** `CumulativeRegretNumber__c`
- **Data Type:** DOUBLE
- **Description:** Total expected reward left on the table by not always picking the globally best-performing arm; computed as sum of PullCount(arm) times (best_rate minus arm_rate) across all arms

### DataSourceId

- **Field API Name:** `DataSourceId__c`
- **Data Type:** TEXT
- **Description:** A unique, system-generated identifier for this record.

### DataSourceObjectId

- **Field API Name:** `DataSourceObjectId__c`
- **Data Type:** TEXT
- **Description:** A unique, system-generated identifier for this record.

### DcsnOptmztnIdentifier

- **Field API Name:** `DcsnOptmztnIdentifier__c`
- **Data Type:** TEXT
- **Description:** Identifier of the decision option configuration this cycle summary belongs to

### EvaluationCycleIdentifier

- **Field API Name:** `EvaluationCycleIdentifier__c`
- **Data Type:** TEXT
- **Description:** UUID identifying the specific evaluation cycle (Spark job run) that produced this record

### EvaluationDateTm

- **Field API Name:** `EvaluationDateTm__c`
- **Data Type:** DATETIME
- **Description:** Timestamp when this evaluation cycle was computed

### Id

- **Field API Name:** `Id__c`
- **Data Type:** TEXT
- **Description:** Primary key composed from DcsnOptimizerConfigurationId + EvaluationCycleId

### InternalOrganizationId

- **Field API Name:** `InternalOrganizationId__c`
- **Data Type:** TEXT
- **Description:** A unique, system-generated identifier for this record.

### KQ_Id

- **Field API Name:** `KQ_Id__c`
- **Data Type:** TEXT
- **Description:** Key Qualifier system field for Id

### MetricName

- **Field API Name:** `MetricName__c`
- **Data Type:** TEXT
- **Description:** Goal metric this cycle summary is computed for taken from the recommender configuration

### OptionCount

- **Field API Name:** `OptionCount__c`
- **Data Type:** DOUBLE
- **Description:** Number of distinct options (arms) active in this evaluation cycle

### TotalCumulativeRewardNumber

- **Field API Name:** `TotalCumulativeRewardNumber__c`
- **Data Type:** DOUBLE
- **Description:** Sum of CumulativeRewardNumber across all arms in the cycle; represents total attributed engagement value for the decision option

### TotalExplorationRatePct

- **Field API Name:** `TotalExplorationRatePct__c`
- **Data Type:** PERCENT
- **Description:** Overall exploration rate across all arms (total explorer count / total pull count) for the cycle

### TotalPullCount

- **Field API Name:** `TotalPullCount__c`
- **Data Type:** DOUBLE
- **Description:** Total number of requests served across all arms in this evaluation cycle

### WeightedAvgRewardRateNbr

- **Field API Name:** `WeightedAvgRewardRateNbr__c`
- **Data Type:** DOUBLE
- **Description:** Pull-count-weighted average of EmpiricalRewardRate across all arms: sum(PullCount times Rate) / sum(PullCount)