OptRL Warehouse Dashboard

Configuration

Basic

Advanced

Warehouse Grid World

▼

Width

Height

Start X

Start Y

Drop-off X

Drop-off Y

Capacity Limit

Max Steps

Shelves List

Stock Portfolio Config

▼

Strategy Presets 1-Click Setup

📈 Core Balanced Default

AAPL · MSFT · SPY

40% single-asset cap with active cash cushion for systematic daily rebalancing.

+5.0% Target 4.0% Max DD

⚡ Tech Momentum Growth

AAPL · MSFT · NVDA

High-conviction tech leaders with dynamic beta allocation and trend momentum.

+8.0% Target 6.0% Max DD

🛡️ Capital Preservation Defensive

SPY · AAPL

Conservative allocation prioritizing downside shield and capital preservation.

+4.0% Target 2.5% Max DD

Initial Capital ($) $100,000

Bar Interval

Max Allocation 40%

Target Profit % Goal

Max Drawdown % Risk

Transaction Fee 0.10% (10 bps)

🔍 Custom Watchlist & Universe Screener ▶

Watchlist Symbols (Search or Comma-Separated)

Canonical Store: Checking sync status...

Market Universe & Rolling Momentum Monthly Insights

The environment calculates rolling relative strength, sector performance, and capital rotation alpha every month. Curate your watchlist from outperforming leaders (for momentum) or oversold laggards (for mean-reversion) to ground reinforcement learning policies.

Curate Watchlist:

Sector Performance:

Loading universe tickers and monthly momentum insights...

⚙️ Advanced Algorithm & Execution Knobs ▶

Episode Length Bars per episode

Walk-Forward Sampling Window evaluation mode

Top-K Active Assets 0 = All assets funded

Chunk Overlap % Window overlap (0 - 90%)

Continuing Capital Rollover Compound portfolio equity across continuous episodes

Goal-Conditioned Policy (GCRL) Condition neural policy on explicit target profit goals

Early Profit Lock-In Settlement Terminate episode early once target profit is attained

Alpha Vantage Key Free Tier (.env)

Auto-reuses saved key.

Reward Formulation

Historical Limit (Days)

Turnover Penalty 0.0010

Reward Scaling 1.0

Donut Shop Config

▼

Horizon (Days)

Lead Time (Days)

Service Weight (fill-rate priority) 0.4

One generalist agent is trained once, then replayed across all three shops (Downtown · Suburban · Airport) on Play Replay.

Training & Algorithm

▼

Algorithm

Training Timesteps 300k

Eval Episodes Evaluation episodes averaged for KPIs

Seed

Execution Mode

Hyperparameters

Idle

Grid World Replay

10 × 8 Grid

Portfolio Value $100,000.00 $0.00 (+0.00%)

Liquid Cash $100,000.00 100.0% Reserve

Capital Exposure 0.0% $0.00 Invested

Drawdown Shield SAFE 0.0% / 4.0% Max DD

Target Profit Goal 0.0% Goal: +5.0%

ADVISORY

Portfolio Manager Advisory Desk

Quantitative Policy Guidance · Real-Time Execution Signals

Advisor Mode: Model Suggestions (Non-Custodial)

Market Leaders (Outperformers)

Awaiting analysis...

Market Laggards (Underperformers)

Awaiting analysis...

Next-Day Tactical

Immediate Session Execution

Awaiting model simulation or recommendation step...

Next-Month Strategic

Allocation & Capital Compounding

Awaiting model simulation or recommendation step...

Cash Reserve

$100,000.00

100.0% Reserve

Invested Capital

$0.00

0.0% Equity

Target Profit Progress

0.0%

Goal: 5.0%

Drawdown Risk

SAFE (0.0%)

Limit: 4.0%

Awaiting model simulation or recommendation step...

Asset & Trend	Action	Price	Allocation Drift	Order Delta	Notional ($)	Conviction	Tactical Outlook & Rationale
No suggestions generated yet. Click Play Replay or query recommendation.

ALLOCATION

Portfolio Asset Allocation & Holdings

Continuous Weight Distribution · Drift from RL Policy Targets

0 Assets Tracked

Asset Allocation Distribution 100.0% Tracked

Cash 100%

Cash: 100.0%

Symbol	Spot Price	Shares Owned	Current Value	Current Weight	Target Weight	Drift Delta
No active portfolio positions yet. Run simulation or generate advisory signals.

LEDGER

Accounting Ledger

Period-by-period returns, fees, and order counts

Period	Date	Open	Close	PnL	Return	Fees	Buys	Sells	Holds
No daily records yet. Click Play Replay to simulate.

SIGNALS

Executed Trade Signals

0 trades

Time	Symbol	Signal	Shares	Price	Notional
No trade signals executed yet.

Now Managing: - idle - Day: 0 Run Profit: $0.00 Fill Rate: -

Train a model, then press Play Replay: the agent works all three shops at once.

Inventory Movement: Replay Timeline

0 days

On-hand Demand Sold summed across all three shops

Daily Ledger

0 rows

Shop	Day	Date	Demand	Sold	Lost	Expired	Ordered	Profit

The impact you just watched

ML predicts. DL predicts better. RL decides.

Profit · 3 shops

Fill rate

Waste

Sell-through

	Machine Learning	Deep Learning	Reinforcement Learning
the job	Predict demand	Predict demand, better	Decide the orders
you get	“~120 tomorrow”	“118 ± 9”	“Order 96 now: lands in 2 days”
lead time · expiry · capacity	✗ needs a separate rule	✗ needs a separate rule	✓ built into the decision
a promo / shock hits	re-forecast, then re-solve	re-forecast, then re-solve	✓ adapts in the same step
optimizes for	forecast accuracy	forecast accuracy	profit · service · waste

ML / DL Predict ▍ stops here OptRL Predict → Decide → Act → Observe ↺

Optimization & adaptation live on the action, not the forecast: a prediction can't tell you how much to order under a 2-day lead time, a shelf that spoils, and capacity shared across products. In this playground, learning to act took profit from $8k → $59.5k (7.4×) and cut waste 57% → 21%. Industry: RL trims inventory cost 15-30% and waste 30-45% vs classical / forecast baselines.

Speed:

Step Trace Feed Clear

System ready. Click Train Model to begin.

KPI & Metrics

Ep: 0

0.0%

Completion Rate

delivered / total

0.0%

Efficiency

delivered / picked

Delivered

total items dropped

Pickups

items lifted from shelves

Drop-offs

items placed at drop-off

Model Performance Rating

No training completed yet

N/A

Training Reward Curve

Mean Return: -

Raw Smoothed len: - · fps: -

Session History

🏆 RLX Board ↗

No sessions found for this user id.

Configuration

Warehouse Grid World

Shelves List

Reward Weights

Stock Portfolio Config

Donut Shop Config

Advanced Inventory Settings

Training & Algorithm

Hyperparameters

Grid World Replay

Portfolio Manager Advisory Desk

Portfolio Asset Allocation & Holdings

Portfolio Value Replay Timeline

Accounting Ledger

Executed Trade Signals

Inventory Movement: Replay Timeline

Daily Ledger

ML predicts. DL predicts better. RL decides.

KPI & Metrics

Training Reward Curve

Session History

Configuration

Warehouse Grid World

Shelves List

Reward Weights

Stock Portfolio Config

Donut Shop Config

Advanced Inventory Settings

Training & Algorithm

Hyperparameters

Grid World Replay

Portfolio Manager Advisory Desk

Portfolio Asset Allocation & Holdings

Portfolio Value Replay Timeline

Accounting Ledger

Executed Trade Signals

Inventory Movement: Replay Timeline

Daily Ledger

ML predicts. DL predicts better. RL decides.

KPI & Metrics

Training Reward Curve

Session History

Sign In to OptRL

Manage API Keys (Alpha Vantage)

Submit to RLX Leaderboard