matlab-solve-optimization

Name: matlab-solve-optimization
Author: matlab

815stars105forks

How to install matlab-solve-optimization?

matlab/matlab-agentic-toolkit/matlab-solve-optimization

$npx -y skills add matlab/matlab-agentic-toolkit --skill matlab-solve-optimization

Installs into the current project.

Summary

Use when writing, solving, or debugging MATLAB optimization code — formulating problems (optimproblem, optimvar, fcn2optimexpr), selecting and configuring solvers (fmincon, linprog, quadprog, intlinprog, lsqnonlin, ga, surrogateopt, optimoptions), or validating results (exitflag,

Files · 1

View on GitHub

SKILL.md

1	# MATLAB Optimization Workflow
2
3	Guide the full optimization lifecycle: classify the problem, formulate it, select and configure a solver, and validate the results.
4
5	## When to Use
6
7	- User is defining an optimization problem in MATLAB (variables, objectives, constraints)
8	- User asks about `optimproblem`, `optimvar`, `optimconstr`, `optimexpr`, or `fcn2optimexpr`
9	- User is selecting or configuring a solver (`optimoptions`, algorithm choice, tuning)
10	- User is interpreting results, debugging convergence, or checking exitflags
11	- User is deciding between problem-based and solver-based approaches
12	- User is writing optimization code with for-loops over decision variables or constraints
13
14	## When NOT to Use
15
16	- User is asking to solve a problem that doesn't require numerical optimization solvers (e.g., finding the minimum value in an array or table)
17	- User is working with non-optimization MATLAB code (data analysis, plotting, signal processing)
18	- User is using a third-party optimization toolbox (not MathWorks)
19	- User is solving symbolic equations with `solve(eqns, vars)`, ODE systems, or linear system solves (`A\b`)
20
21	---
22
23	## Stage 1: Classify & Formulate
24
25	### 1.1 Classify the Problem
26
27	Before formulating, identify the problem class — it determines which solver to use, what guarantee you can promise (global vs local), and whether a domain-specific tool should replace the generic path.
28
29	See [references/classify.md](references/classify.md) for the class→solver→guarantee table, convexity quick-checks, and "hidden easier class" heuristics. Key actions:
30	- Check if a purpose-built domain tool exists before falling back to `optimproblem`
31	- Watch for hidden easier classes (sum-of-squares disguised as NLP, linear structure missed)
32	- For QPs, check `eig(H)` — nonconvex QPs cannot use `quadprog` reliably
33	- Watch for hidden nonsmoothness: `max`, `min`, `abs`, `sort`, `if`/branching, or norms other than squared-2-norm
34
35	### 1.2 Choose Approach
36
37	Use problem-based by default for readable definitions, N-D modeling, and every LP, QP, conic, and mixed-integer problem (unless coefficients are already in matrix-vector form). Problem-based provides automatic differentiation and is less error-prone.
38
39	Even when AD is blocked (e.g., `ode45` in the objective), `fcn2optimexpr` can still wrap the function as a black-box — problem-based remains useful.
40
41	Only fall back to solver-based when one of these applies:
42
43	\| Use solver-based when... \| Reason \|
44	\|---\|---\|
45	\| Trivial mapping to solver API — one vector `x`, pre-coded objective with exact gradients/Hessian \| No benefit from abstraction; solver-based is direct \|
46	\| Overhead of building problem-based expressions dominates computation \| Avoid tracing/transformation overhead \|
47	\| Need a solver feature problem-based doesn't expose (`CheckpointFile`, exact Hessians, custom `OutputFcn`) \| Only available via solver-based calls \|
48	\| C code generation for embedded deployment is required \| Problem-based does not support codegen \|
49
50	Converting between approaches: `prob2struct(prob)` converts problem-based to solver-based form for deployment or performance.
51
52	References:
53	- Problem-based: [references/problem-based-guide.md](references/problem-based-guide.md)
54	- Solver-based (class→solver mapping): [references/classify.md](references/classify.md)
55
56	### 1.3 Formulate the Problem
57
58	Problem-based canonical template:
59
60	```matlab
61	% 1. Define decision variables
62	x = optimvar("x", N, LowerBound=lb, UpperBound=ub);
63
64	% 2. Create problem
65	prob = optimproblem("Objective", sum(x,"all"));
66
67	% 3. Add constraints
68	prob.Constraints.linear = A*x <= b;
69	prob.Constraints.nonlinear = fcn2optimexpr(@myNonlinFcn, x) <= rhs;
70
71	% 4. Set initial guess (must be struct with field names matching optimvar names)
72	x0.x = initialValues;
73
74	% 5. Solve
75	[sol, fval, exitflag, output] = solve(prob, x0);
76	```
77
78	Solver-based key differences:
79	- Initial guess is a numeric vector, not a struct
80	- You manage variable indexing manually (flat vector `x`)
81	- Supply gradients manually for best performance (`SpecifyObjectiveGradient=true`)
82	- Linear/quadratic solvers require explicit coefficient matrices
83
84	### 1.4 Validate at the Start Point
85
86	Before calling any solver, evaluate the objective and constraints at `x0` to catch sign/size/NaN errors early:
87
88	```matlab
89	% Problem-based
90	fval0 = evaluate(prob.Objective, x0);
91	assert(isfinite(fval0), 'Objective is not finite at x0');
92	infeas0 = infeasibility(prob.Constraint

People who liked this also viewed

6 picks

Type

matlab-solve-optimization

815stars105forks

View on GitHub

How to install matlab-solve-optimization?

matlab/matlab-agentic-toolkit/matlab-solve-optimization

$npx -y skills add matlab/matlab-agentic-toolkit --skill matlab-solve-optimization

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# MATLAB Optimization Workflow
2
3	Guide the full optimization lifecycle: classify the problem, formulate it, select and configure a solver, and validate the results.
4
5	## When to Use
6
7	- User is defining an optimization problem in MATLAB (variables, objectives, constraints)
8	- User asks about `optimproblem`, `optimvar`, `optimconstr`, `optimexpr`, or `fcn2optimexpr`
9	- User is selecting or configuring a solver (`optimoptions`, algorithm choice, tuning)
10	- User is interpreting results, debugging convergence, or checking exitflags
11	- User is deciding between problem-based and solver-based approaches
12	- User is writing optimization code with for-loops over decision variables or constraints
13
14	## When NOT to Use
15
16	- User is asking to solve a problem that doesn't require numerical optimization solvers (e.g., finding the minimum value in an array or table)
17	- User is working with non-optimization MATLAB code (data analysis, plotting, signal processing)
18	- User is using a third-party optimization toolbox (not MathWorks)
19	- User is solving symbolic equations with `solve(eqns, vars)`, ODE systems, or linear system solves (`A\b`)
20
21	---
22
23	## Stage 1: Classify & Formulate
24
25	### 1.1 Classify the Problem
26
27	Before formulating, identify the problem class — it determines which solver to use, what guarantee you can promise (global vs local), and whether a domain-specific tool should replace the generic path.
28
29	See [references/classify.md](references/classify.md) for the class→solver→guarantee table, convexity quick-checks, and "hidden easier class" heuristics. Key actions:
30	- Check if a purpose-built domain tool exists before falling back to `optimproblem`
31	- Watch for hidden easier classes (sum-of-squares disguised as NLP, linear structure missed)
32	- For QPs, check `eig(H)` — nonconvex QPs cannot use `quadprog` reliably
33	- Watch for hidden nonsmoothness: `max`, `min`, `abs`, `sort`, `if`/branching, or norms other than squared-2-norm
34
35	### 1.2 Choose Approach
36
37	Use problem-based by default for readable definitions, N-D modeling, and every LP, QP, conic, and mixed-integer problem (unless coefficients are already in matrix-vector form). Problem-based provides automatic differentiation and is less error-prone.
38
39	Even when AD is blocked (e.g., `ode45` in the objective), `fcn2optimexpr` can still wrap the function as a black-box — problem-based remains useful.
40
41	Only fall back to solver-based when one of these applies:
42
43	\| Use solver-based when... \| Reason \|
44	\|---\|---\|
45	\| Trivial mapping to solver API — one vector `x`, pre-coded objective with exact gradients/Hessian \| No benefit from abstraction; solver-based is direct \|
46	\| Overhead of building problem-based expressions dominates computation \| Avoid tracing/transformation overhead \|
47	\| Need a solver feature problem-based doesn't expose (`CheckpointFile`, exact Hessians, custom `OutputFcn`) \| Only available via solver-based calls \|
48	\| C code generation for embedded deployment is required \| Problem-based does not support codegen \|
49
50	Converting between approaches: `prob2struct(prob)` converts problem-based to solver-based form for deployment or performance.
51
52	References:
53	- Problem-based: [references/problem-based-guide.md](references/problem-based-guide.md)
54	- Solver-based (class→solver mapping): [references/classify.md](references/classify.md)
55
56	### 1.3 Formulate the Problem
57
58	Problem-based canonical template:
59
60	```matlab
61	% 1. Define decision variables
62	x = optimvar("x", N, LowerBound=lb, UpperBound=ub);
63
64	% 2. Create problem
65	prob = optimproblem("Objective", sum(x,"all"));
66
67	% 3. Add constraints
68	prob.Constraints.linear = A*x <= b;
69	prob.Constraints.nonlinear = fcn2optimexpr(@myNonlinFcn, x) <= rhs;
70
71	% 4. Set initial guess (must be struct with field names matching optimvar names)
72	x0.x = initialValues;
73
74	% 5. Solve
75	[sol, fval, exitflag, output] = solve(prob, x0);
76	```
77
78	Solver-based key differences:
79	- Initial guess is a numeric vector, not a struct
80	- You manage variable indexing manually (flat vector `x`)
81	- Supply gradients manually for best performance (`SpecifyObjectiveGradient=true`)
82	- Linear/quadratic solvers require explicit coefficient matrices
83
84	### 1.4 Validate at the Start Point
85
86	Before calling any solver, evaluate the objective and constraints at `x0` to catch sign/size/NaN errors early:
87
88	```matlab
89	% Problem-based
90	fval0 = evaluate(prob.Objective, x0);
91	assert(isfinite(fval0), 'Objective is not finite at x0');
92	infeas0 = infeasibility(prob.Constraint

People who liked this also viewed

6 picks

Type