Qavis Inc.

Overview

Qavis builds a hybrid quantum-AI optimization platform currently serving logistics (vehicle routing, delivery scheduling). We are expanding into manufacturing, where Job-Shop Scheduling (JSSP) is a core challenge: N jobs must be processed across M machines, each operation requiring a specific machine for a specific duration, with precedence and no-overlap constraints, minimizing makespan. We need a complete mathematical formulation of JSSP as a binary quadratic optimization problem suitable for quantum and hybrid solvers.

Problem Statement

The key research questions are:

1. How should decision variables, objective, and constraints be encoded as quadratic penalty terms, and which variable encoding (time-indexed, position-based, order-based) yields the best structure?
2. How should penalty coefficients be chosen — can we derive theoretical bounds and sensitivity guidelines?
3. How can large instances be decomposed into smaller sub-problems while preserving solution quality? The project is purely mathematical - no quantum hardware or production software is required.

Deliverables are a formal formulation document, penalty analysis, decomposition strategy, and an M2PI final report. Qavis provides benchmark datasets, domain context, and optional IBM simulator access for interested team members.