isso

README

isso - Iterative Sampling Schedule Optimization

isso is a Go library and CLI app for optimizing sampling schedules under constrained capacity and with potential sample re-use.

CLI usage

Run the included examples like this...

The default test problem:

go run ./cmd/isso -i data/problem.json

A pareto optimization example:

go run ./cmd/isso -i data/pareto.json --pareto --format fitness

See folder data for problem definition examples.

License

This project is distributed under the MIT license.

Documentation

Index

• type Action
• type ActionDef
• type Comparator
• type Evaluator
• type Matrix
• type Problem
  • func NewProblem(problem ProblemDef) Problem
• type ProblemDef
• type Requirement
• type Solution
  • func (s *Solution[F]) ToCSV(index int, sep string) string
  • func (s *Solution[F]) ToList() string
  • func (s *Solution[F]) ToTable() string
• type Solver
  • func NewSolver[F comparable](evaluator Evaluator[F], comparator Comparator[F]) Solver[F]
  • func (s *Solver[F]) Solve(problem *Problem) ([]Solution[F], bool)

Types

type Action

type Action struct {
	Subject       string
	Matrix        string
	Reuse         string
	Time          int
	Samples       int
	TargetSamples int
}

Action definition.

type ActionDef

type ActionDef struct {
	Subject       subject
	Matrix        matrix
	Reuse         subject
	Time          int
	Samples       int
	TargetSamples int
}

ActionDef for internal use, using no strings. It needs to be public as it is used in fitness evaluators.

type Comparator

type Comparator[F any] interface {
	Compare(a, b F) int
	IsPareto() bool
}

Comparator interface or comparing fitness values.

type Evaluator

type Evaluator[F any] interface {
	Evaluate(s []ActionDef) F
}

Evaluator interface or deriving fitness from a solution.

type Matrix

type Matrix struct {
	Name     string
	CanReuse []string
}

Matrix definition.

type Problem

type Problem struct {
	// contains filtered or unexported fields
}

Problem definition.

func NewProblem

func NewProblem(problem ProblemDef) Problem

NewProblem creates a new problem definition.

type ProblemDef

type ProblemDef struct {
	Matrices     []Matrix
	Capacity     []int
	Requirements []Requirement
}

type Requirement

type Requirement struct {
	Subject string
	Matrix  string
	Times   []int
	Samples int
}

Requirement definition.

type Solution

type Solution[F any] struct {
	Fitness F
	Actions []Action
}

Solution, translated back to using strings for subject and matrix.

func (*Solution[F]) ToCSV

func (s *Solution[F]) ToCSV(index int, sep string) string

ToCSV formats the solution as a CSV table.

func (*Solution[F]) ToList

func (s *Solution[F]) ToList() string

ToList formats the solution as list for printing.

func (*Solution[F]) ToTable

func (s *Solution[F]) ToTable() string

ToTable formats the solution as a table for printing.

type Solver

type Solver[F comparable] struct {
	// contains filtered or unexported fields
}

Solver for optimization.

func NewSolver

func NewSolver[F comparable](evaluator Evaluator[F], comparator Comparator[F]) Solver[F]

NewSolver creates a new solver for a given fitness function.

func (*Solver[F]) Solve

func (s *Solver[F]) Solve(problem *Problem) ([]Solution[F], bool)

Solve the given problem.