Article

2026-03-05 Science & Technology

Archived material is restricted to Rantburg regulars and members. If you need access email fred.pruitt=at=gmail.com with your nick to be added to the members list. There is no charge to join Rantburg as a member.

Posted by badanov 2026-03-05 00:00|| E-Mail|| Front Page|| [1779 views ] Top

1 Been happening since scenario modeling began.

[AI]
'Excel Solver Chooses Nukes'

The phrase "Excel Solver chooses nukes" appears to be a metaphorical or humorous way of expressing that Solver is selecting an extreme or overly aggressive solution—perhaps one that seems disproportionate or unintended—when optimizing a model.

In practical terms, this could happen due to:

- **Poorly defined constraints**: If constraints are too restrictive or conflicting, Solver may find a solution that maximizes or minimizes the objective function in an extreme way, such as allocating all resources to one option (e.g., "choosing nukes" over other alternatives).

- **Non-linear or discontinuous models**: When the objective function or constraints involve non-linear relationships (e.g., binary variables multiplied by continuous ones), Solver may struggle and converge on an unexpected or suboptimal extreme solution.

- **Starting point influence**: Solver uses gradient-based methods (like GRG Nonlinear), which can lead to different results depending on the initial values. A poor starting point might push the solution toward an extreme outcome.

- **Unintended variable behavior**: If variables are not properly constrained (e.g., allowing negative values when not intended), Solver may exploit this to achieve a "better" objective value in a counterintuitive way.

To address this:
- **Review constraints** and ensure they reflect real-world limits.
- Use **"Make Unconstrained Variables Non-Negative"** in Solver Options if negative values are not allowed.
- Try **different solving methods** (e.g., Simplex LP for linear problems, or OpenSolver for more robust handling of complex models).
- Set **integer or binary constraints** explicitly when needed.
- Test with **different starting values** to check solution consistency.

If the model is complex, consider **linearizing** non-linear components (e.g., avoid multiplying continuous and binary variables directly) to improve Solver reliability.

Posted by Skidmark 2026-03-05 07:26|| 2026-03-05 07:26|| Front Page Top

2 China ramps up 'high stakes' tech race with US as economic imbalances deepen

Posted by Skidmark 2026-03-05 15:31|| 2026-03-05 15:31|| Front Page Top

23:55 Skidmark
22:58 trailing wife
22:54 trailing wife
22:52 trailing wife
22:50 trailing wife
22:36 trailing wife
21:39 Super Hose
21:33 Super Hose
21:31 Super Hose
21:29 Super Hose
20:57 Enver Omump7759
20:24 Woodrow
20:11 Glavirong Unolurt5009
19:27 Procopius2k
19:15 Enver Omump7759
18:35 Frank G
18:23 SteveS
18:22 Procopius2k
18:06 Lord Garth
18:03 Frank G
18:00 SteveS
17:55 swksvolFF
17:51 trailing wife
17:47 trailing wife
Merry-Go-Blog