Engineering portfolio selection
Choose the combination—not just the highest-ranked projects.
A list of attractive initiatives is not a feasible portfolio. DAOSoft helps teams examine which combination fits the available resources and supports the program’s priorities.
Define the portfolio the program can actually deliver.
Set candidate initiatives, expected benefits, costs and resource limits. Model prerequisites, mutually exclusive choices and other selection rules before comparing combinations.
Test the combination against changing conditions.
Explore reduced funding, supplier delay, cost escalation or changing priorities. Compare the selected portfolio with each scenario’s best feasible result. A robust choice need not rank first in every scenario.
Keep the models used to estimate benefits and performance visible, so decision-makers can examine what drives the result.
Worked example · synthetic data
One feasible mix. Four sets of priorities.
The selected four-investment portfolio protects the lowest relative result across these scenarios. It is not the highest-scoring portfolio in every scenario.
The scenarios change benefit priorities. Costs, staffing limits and selection rules stay fixed.
Modeled decision value relative to each scenario’s best feasible mix
Balanced program
95.4%Severe winter
81.0%Supplier delay
100.0%Cost squeeze
90.9%Each row is scaled to that scenario’s best feasible mix (100%). It is not a probability or a financial return.
Selected investment mix
- Winter test capacity
- Pack supplier qualification
- Controls release
- Fleet evidence pipeline
The actual tradeoff: compared with the balanced-scenario optimum, this mix uses $0.5M more and two additional staff-months. It gives up three balanced-scenario value points (62 rather than 65), but raises the weakest relative result from 75.0% to 81.0%.
A deterministic calculation over seven candidate investments.
Inspect the method and limits
All 128 subsets were enumerated; 16 met the modeled resource and selection rules. The decision rule maximizes the weakest ratio to a scenario-specific optimum, then breaks ties by lower maximum regret and cost. Benefits are constructed additive decision scores; resources are deterministic. No scenario probabilities or later recourse are modeled.
For this demonstration, scores are constructed cardinal points: zero means no modeled benefit contribution, and candidate contributions add in common scenario units. This is an explicit interpretation contract—not an elicited customer utility model. Ratios would not be justified for arbitrary ordinal ranks. No primary-score tie affects the selected result.
Inspect the selected portfolio and scenario results · Read the original model card
Examine a proposed portfolio—including human adjustments.
Review a model-generated portfolio or assemble a manual candidate and check it against the modeled constraints. Record the reason for any departure from the analytical recommendation.
Leave the review with a selection and its basis.
Bring the selected combination, resource use, scenario exposure and rationale into a Decision Review Package. Decision-makers can inspect the assumptions and the consequences of choosing a different mix.
Explore a proposed mix, its constraints and the effect of a different choice.
Bring a candidate investment mix. Explore its constraint check, a human adjustment and the recorded rationale.