Platform & Deployment

Flexible planning for repeatable, governed execution

Assemble dataset, model configuration, and engine configuration bundles to make scope and behavior explicit and reproducible, with ultra visibility into results. Deploy as a dedicated environment when tailored operational governance is required.

Case workspace and system flexibility

PentaTorch is organized around cases. Each case is a controlled, reproducible planning setup that teams can configure, run, compare, and review without hidden assumptions. A case combines a Dataset instance, a Model Configuration bundle, and an Engine Configuration bundle, so scope and behavior are selected deliberately and remain repeatable across runs. This structure supports controlled experimentation and side by side comparisons with clear attribution to what changed. With ultra visibility and flexible filtering and drill-down views, teams can examine outcomes at the right level of detail to validate feasibility and support handoffs. Engine choice and run controls stay inside the case definition, so scaling up keeps governance and comparability intact. Dataset ingestion is schema validated so inputs remain consistent across runs and environments. Deployment can be provided as a dedicated environment when tailored operational governance is required.

View case structure details
Case-based workflow
Create, clone, and compare cases as first-class objects—so iteration is safe, governed, and reviewable.
Bundle-driven configuration
Assemble a case by combining Dataset, Model Config, and Engine Config bundles—making choices explicit and repeatable.
Query-first visibility
Ultra visibility into every case and run through filtering and drill-down views. Trace outcomes through the underlying plan tables and support feasibility checks and handoffs without manual rework.
Engines designed for scale
Select engines appropriate to problem size and planning cadence. Because engine choice and run controls are part of the case definition, performance tuning and algorithm selection stay governed—and comparisons remain consistent across runs.
Integration
Database-native case storage with schema-based ingestion and validation. Designed for API-first integration so datasets, runs, and outputs connect cleanly to surrounding planning and execution systems.
Deployment options
Supports single-tenant deployments for organizations that require isolation and controlled access, with environment-separated configuration aligned to enterprise governance. Implementation support is available for initial setup, training/enablement, and ongoing maintenance.
Case structure details
Case definition
Each case is built from three components: a Dataset instance (master data and scenario definitions), a Model Configuration bundle (scope, objective mode, constraints and business rules), and an Engine Configuration bundle (engine choice and run controls). This keeps modeling and engine features selectable via configuration, without rewriting logic.
Why this design holds up in practice
Reproducible results
Cases capture the full configuration so outcomes can be trusted, revisited, and reviewed consistently.
Controlled experimentation
Swap bundles intentionally to change the dataset, the model options, or the engine settings while keeping comparisons apples to apples.
Case-by-case comparisons
Compare cases side by side with clear attribution to configuration differences, so it is clear what changed and why.
What the platform covers
Network footprint
DC selection, supplier selection, and activation/open decisions—defined as part of the case and enforced end-to-end.
Routing, lanes, and connectivity
Inbound and outbound connectivity with lane eligibility rules so plans respect real network constraints.
Transportation mode selection
Mode choice per lane with feasibility and service constraints enforced, aligned to execution reality.
Inbound and outbound flows
Flow quantities across the network with capacity and feasibility enforced across the full structure.
Procurement and sourcing
Sourcing quantities and commercial rules enforced across suppliers, products, and periods.
Production, time-phased inventory, and service
When enabled, production planning and time-phased inventory with service enforcement across DCs, products, and periods.