Constraint programming

Action
Agent-oriented
Array-oriented
Automata-based
Concept
Concurrent computing
- Relativistic programming
Data-driven
Declarative (contrast: Imperative)
- Constraint
  - Constraint logic
    - Concurrent constraint logic
- Dataflow
  - Flow-based
  - Cell-oriented (spreadsheets)
  - Reactive
- Functional
  - Functional logic
  - Purely functional
- Logic
Dynamic
End-user programming
Event-driven
- Service-oriented
- Time-driven
Expression-oriented
Feature-oriented
Function-level (contrast: Value-level)
Generic
Imperative (contrast: Declarative)
- Literate
- Procedural
Inductive programming
Language-oriented
- Natural language programming
- Discipline-specific
- Domain-specific
- Grammar-oriented
  - Dialecting
- Intentional
Metaprogramming
- Automatic
- Reflective
  - Attribute-oriented
- Homoiconic
- Macro
- Template
  - Policy-based
Non-structured (contrast: Structured)
- Array
Nondeterministic
Parallel computing
- Process-oriented
Point-free style
- Concatenative
Probabilistic
Semantic
Structured (contrast: Non-structured)
- Block-structured
- Modular (contrast: Monolithic)
- Object-oriented
- Recursive
Value-level (contrast: Function-level)

In computer science, constraint programming is a programming paradigm wherein relations between variables are stated in the form of constraints. Constraints differ from the common primitives of imperative programming languages in that they do not specify a step or sequence of steps to execute, but rather the properties of a solution to be found. This makes constraint programming a form of declarative programming. The constraints used in constraint programming are of various kinds: those used in constraint satisfaction problems (e.g. "A or B is true"), those solved by the simplex algorithm (e.g. " $x$ ≤ 5"), and others. Constraints are usually embedded within a programming language or provided via separate software libraries.

Constraint programming can be expressed in the form of constraint logic programming, which embeds constraints into a logic program. This variant of logic programming is due to Jaffar and Lassez, who extended in 1987 a specific class of constraints that were introduced in Prolog II. The first implementations of constraint logic programming were Prolog III, CLP(R), and CHIP.

Instead of logic programming, constraints can be mixed with functional programming, term rewriting, and imperative languages. Programming languages with built-in support for constraints include Oz (functional programming) and Kaleidoscope (imperative programming). Mostly, constraints are implemented in imperative languages via constraint solving toolkits, which are separate libraries for an existing imperative language.

Constraint logic programming

Constraint programming is an embedding of constraints in a host language. The first host languages used were logic programming languages, so the field was initially called constraint logic programming. The two paradigms share many important features, like logical variables and backtracking. Today most Prolog implementations include one or more libraries for constraint logic programming.

The difference between the two is largely in their styles and approaches to modeling the world. Some problems are more natural (and thus, simpler) to write as logic programs, while some are more natural to write as constraint programs.

The constraint programming approach is to search for a state of the world in which a large number of constraints are satisfied at the same time. A problem is typically stated as a state of the world containing a number of unknown variables. The constraint program searches for values for all the variables.

Temporal concurrent constraint programming (TCC) and non-deterministic temporal concurrent constraint programming (MJV) are variants of constraint programming that can deal with time.

Perturbation vs refinement models

Languages for constraint-based programming follow one of two approaches:^[1]

Refinement model: variables in the problem are initially unassigned, and each variable is assumed to be able to contain any value included in its range or domain. As computation progresses, values in the domain of a variable are pruned if they are shown to be incompatible with the possible values of other variables, until a single value is found for each variable.
Perturbation model: variables in the problem are assigned a single initial value. At different times one or more variables receive perturbations (changes to their old value), and the system propagates the change trying to assign new values to other variables that are consistent with the perturbation.

Constraint propagation in constraint satisfaction problems is a typical example of a refinement model, and spreadsheets are a typical example of a perturbation model.

The refinement model is more general, as it does not restrict variables to have a single value, it can lead to several solutions to the same problem. However, the perturbation model is more intuitive for programmers using mixed imperative constraint object-oriented languages.^[2]

Domains

The constraints used in constraint programming are typically over some specific domains. Some popular domains for constraint programming are:

boolean domains, where only true/false constraints apply (SAT problem)
integer domains, rational domains
linear domains, where only linear functions are described and analyzed (although approaches to non-linear problems do exist)
finite domains, where constraints are defined over finite sets
mixed domains, involving two or more of the above

Finite domains is one of the most successful domains of constraint programming. In some areas (like operations research) constraint programming is often identified with constraint programming over finite domains.

All of the above examples are commonly solved by satisfiability modulo theories (SMT) solvers.

Finite domain solvers are useful for solving constraint satisfaction problems, and are often based on arc consistency or one of its approximations.

The syntax for expressing constraints over finite domains depends on the host language. The following is a Prolog program that solves the classical alphametic puzzle SEND+MORE=MONEY in constraint logic programming:

% This code works in both YAP and SWI-Prolog using the environment-supplied
% CLPFD constraint solver library.  It may require minor modifications to work
% in other Prolog environments or using other constraint solvers.
:- use_module(library(clpfd)).
sendmore(Digits) :-
   Digits = [S,E,N,D,M,O,R,Y],     % Create variables
   Digits ins 0..9,                % Associate domains to variables
   S #\= 0,                        % Constraint: S must be different from 0
   M #\= 0,
   all_different(Digits),          % all the elements must take different values
                1000*S + 100*E + 10*N + D     % Other constraints
              + 1000*M + 100*O + 10*R + E
   #= 10000*M + 1000*O + 100*N + 10*E + Y,
   label(Digits).                  % Start the search

The interpreter creates a variable for each letter in the puzzle. The operator ins is used to specify the domains of these variables, so that they range over the set of values {0,1,2,3, ..., 9}. The constraints S#\=0 and M#\=0 means that these two variables cannot take the value zero. When the interpreter evaluates these constraints, it reduces the domains of these two variables by removing the value 0 from them. Then, the constraint all_different(Digits) is considered; it does not reduce any domain, so it is simply stored. The last constraint specifies that the digits assigned to the letters must be such that "SEND+MORE=MONEY" holds when each letter is replaced by its corresponding digit. From this constraint, the solver infers that M=1. All stored constraints involving variable M are awakened: in this case, constraint propagation on the all_different constraint removes value 1 from the domain of all the remaining variables. Constraint propagation may solve the problem by reducing all domains to a single value, it may prove that the problem has no solution by reducing a domain to the empty set, but may also terminate without proving satisfiability or unsatisfiability. The label literals are used to actually perform search for a solution.