Guide
Scope & Limits
SSPLAX builds one linear or mixed-integer model of your system and finds the best feasible plan for the objective you choose. That covers a wide range of operational decisions, though not every problem. The boundaries are worth knowing before investing in a model.
Where SSPLAX fits
A problem fits when the decision is how much or which options, and the relationships between quantities are proportional. That covers more ground than the flow network alone:
Material, work, or money moving through connected stages under capacities, budgets, and delivery requirements. Production lines, supply routes, multi-stage processes.
Splitting a shared pool such as budget, headcount, or machine time across competing uses, or assigning work to the places able to do it.
Which suppliers to use and how much to take from each; which products to make given shared capacity and margins.
Portfolio and packing problems: choose a subset of options, subject to counts, dependencies, exclusions, and a resource ceiling.
The same system across multiple periods, with demand growth, ramp limits, inventory carry-over, backlog, and holding costs.
Weighted scenarios, a check of an already-committed plan, and expected-value or worst-case optimization across the whole set.
Structured rostering and staffing also fit, provided the question is how many people of each type are needed in each period, rather than who works which named shift.
What doesn't fit
These would need mathematics SSPLAX doesn't currently have. They aren't missing controls; they're a different class of problem.
Which job runs first, what time a campaign starts, how changeovers order across a week. SSPLAX plans in aggregate by period. It reports how much happens in each period, not the calendar within it.
How many nurses of each grade per period is in scope. Which named nurse covers Tuesday night is not.
Choosing the order of stops on a route is a sequencing problem, not a flow problem, and needs a dedicated routing engine.
Costs and yields are straight lines, or piecewise-linear across breakpoints you define. Effects that genuinely curve, such as congestion that grows with load, learning curves, or smooth economies of scale, are approximated at best.
SSPLAX optimizes the model you give it. It does not forecast demand or learn rates from history. A wrong capacity produces a confidently wrong plan, which is why input documentation and pressure testing matter.
Two different boundaries
It's worth separating "SSPLAX can't do this" from "there isn't a button for it yet."
Sequencing, routing, exact start times, and general nonlinear optimization. These need new solver capability, so no amount of clever modeling gets you there.
A few capabilities the model and solver fully support do not yet have a dedicated control, most notably scenario-specific resource rates. These can be supplied through a spreadsheet import. See Import Model Data.
If your problem fits, Build a Model walks through the network side, and Decisions & Targets covers the choices that aren't flows.