## MiniZinc Challenge 2016 -- Rules

These are the official rules for the MiniZinc Challenge 2016. Version 2.0.13.

These rules were last updated on 19 May 2016.

### Entrants

The MiniZinc Challenge 2016 will test solvers on problems written in MiniZinc 2.0.13.

Let *name* be the name of the solver system in what follows.

An entrant in the challenge is a constraint solver that is installed in a virtual machine (VM) provided by the organizers. The download link will be provided after registration.

Constraint solvers that have several variants, for example that can alternatively use copying or trailing, may submit one entry per variant although the organizers reserve the right to reject such variations if they not sufficiently interesting, (e.g. multiple copies of the same solver with differing parameters).

Each entrant must provide a gzipped tarball containing the following:

A text file named CLASSES specifying which competition CLASS(es) the entry is to be entered in.

A text file named DESCRIPTION, that contains a short (1-2 pages) description of the system. This should include a list of all authors of the system and their present institutional affiliations. It should also describe any algorithms or data structures that are not standardly used in such systems.

System descriptions will be posted on the MiniZinc Challenge 2016 website.The provided VM with the installed solver can be run by the provided scripts inside the VM as FlatZinc or MiniZinc solver, i.e.,

**FlatZinc Solver**

fzn-exec - an executable file in the VM folder /home/user/entry_data that invokes a FlatZinc solver handling FlatZinc version 2.0.13. If your FlatZinc solver**only**handles XML-FlatZinc version 2.0.13, please contact the organizers.

This executable will be invoked from the command line/scripts as follows:

The argument**fzn-exec**[<options>] file.fzn*file.fzn*is the name of a FlatZinc 2.0.13 model instance to evaluate.**MiniZinc Solver**

An executable file needs to provided in the VM folder /home/user/entry_data that invokes a MiniZinc solver handling MiniZinc version 2.0.13. This executable will be invoked from the command line and/or scripts located at the VM folder /home/user/bin as follows:

The arguments**exec**[<options>] -G <mznlibdir> file.mzn [<data.dzn>]*file.mzn*and*data.dzn*are names of a MiniZinc 2.0.13 model and data file, respectively. The argument*-G <mznlibdir>*points to the location of your solver's MiniZinc library directory containing the redefinition or global constraints files.

*-a***Satisfaction problems**

This causes the solver to search for, and output all solutions.

When this option is not given the solver should search for, and output the first solution.**Optimisation problems**

This causes the solver to search for the first optimal solution, and output all found intermediate solutions and the first optimal solution.

When this option is not given the solver should search for, and output the first optimal solution.

*-f*

When invoked with this option the solver is free to ignore any specified search strategy.*-p <n>*

When invoked with this option the solver is free to use multiple threads and/or cores during search. The argument*n*specifies the number of cores that are available.

Execution of solvers must not require root access.

Any solver-specific definitions of the global constraints in the MiniZinc library in the VM directory

*/home/user/entry_data/mzn-lib*.

This directory may also contain a file named*redefinitions.mzn*that contains redefinitions of FlatZinc built-ins required by the solver.

The gzipped tar-ball must be made accessible for download for the organizer and the submitter must send an email to the organizer describing how to download the modified VM.

The organizers will make reasonable efforts to run each system, including communication with the submitters of the system in case of difficulties. Nevertheless, the organizers reserve the right to reject an entrant if its process proves overly difficult.

The results will be announced at CP2016. Entrants are encouraged to physically attend CP2016, but are not required to in order to participate or win.

There will be at most five competition CLASSES depending on how many solvers are entered in each of them:

- FD search: solvers in this class must follow the search strategy defined in the problem, they will be disqualified if there is evidence that they do not follow the search strategy.
- Free search: solvers in this class are free to ignore the search strategy. All FD search solvers will be automatically entered in this class too.
- Parallel search: solvers in this class are free to use multiple threads or cores to solve the problem. All entrants in the free search class will be automatically entered in this class too, but they will be run in a single threaded mode.
- Open class: This class allows the usage of portfolio solvers. Solvers in this class are free to use multiple threads or cores to solve the problem. All entrants in the parallel search class will be automatically entered in this class too.

The CLASSES file included in the entry must specify which competition CLASS(es) the entry is to be entered in.

### Problem Format

The problem format will be MiniZinc 2.0.13.

There will be some restrictions on the problems tested in MiniZinc challenge.

- No floats may be involved in any model. This is to avoid accuracy differences and simplify entry requirements.
- No variable sets of integers may be used in any model. This is to simplify entry requirements. Not even implicit var sets of int, e.g. this is forbidden:
`array[1..3] of set of 1..3: a = [{1,2}, {3}, {1,3}]; var 1..3: i; constraint card(a[i]) > 1;`

- In order to facilitate local search entrants, ideally a model should wrap symmetry breaking constraints in a predicate “symmetry_breaking_constraint” e.g.,

and redundant constraints in a predicate “redundant_constraint”, e.g.,`var 0..100: x; var 0..100: y; constraint x + y < 144; constraint symmetry_breaking_constraint(x <= y);`

`array[1..4] of var 0..20: start; array[1..4] of int: duration = [3, 4, 6, 7]; array[1..4] of int: usage = [6, 3, 5, 3]; constraint cumulative(start, duration, usage, 10); constraint redundant_constraint(start[1] + dur[1] <= start[3] \/ start[3] + dur[3] <= start[1]);`

- Each solve item must be annotated with a search strategy, such that fixing all the variables appearing in the search strategy would allow a value propagation solver to check a solution. For example,

is correct but not`var 1..5: x; var 1..5: y; var 1..5: z; constraint x <= y /\ y <= z; solve :: int_search([x, y, z], input_order, indomain_min, complete) satisfy;`

even though most FD solvers would know the second was satisfiable.`solve :: int_search([x,z], input_order, indomain_min, complete) satisfy;`

- Search annotations will be restricted to
**bool_search**,**int_search**and**seq_search**.

For**bool_search**and**int_search**only the following parameters (where applicable) will be used:- variable choice:
- input_order
- first_fail (variable with smallest domain size)
- anti_first_fail (variable with largest domain size)
- smallest (variable with smallest minimal value)
- largest (variable with largest maximum value)

- value choice: [examples for domain {1,3,4,18}]
- indomain_min (x <= 1; x > 1)
- indomain_max (x >= 18; x < 18)
- indomain_median (x = 3 ; x != 3)
- indomain_split (x <= (1+18)/2 ; x > (1+18)/2 )
- indomain_reverse_split (x > (1+18)/2 ; x <= (1+18)/2 )

- search style
- complete

will first label`var 1..5: x; var 1..10: objective; constraint x > 1 -> objective > 7; constraint x = 1 -> objective < 3; solve :: int_search([x, objective], first_fail, indomain_min, complete) maximize objective;`

`x = 1`

and find maximum value`objective = 2`

eventually on backtracking to the choice`x = 1`

, then setting`x >= 2`

in most FD solvers will have domains for`x :: 2..5`

and`objective :: 8..10`

and this time`objective`

will be selected as the next variable to label. A full specification of the available choices is given in the FlatZinc 1.6 specification. - variable choice:
- The objective variable must be called
**objective**in optimisation problems, e.g. see previous example.

### Output Requirements

Output from entries must conform to the FlatZinc 1.6 specification. For optimization problems, if the time limit is exceeded before the final solution is printed then the last complete approximate solution printed will be considered to be the solution for that entry. Note that is important that entries flush the output stream after printing each approximate solution.

### Execution Environment

During the MiniZinc Challenge 2016 all VMs will run on machines with the following specification:

- Operating System: Ubuntu 14.04.4 LTS
- Processor(s): i7 3770 @ 3.40GHz (8 logical cores)
- Memory: 16 Gb
- VirtualBox 5.0.20r106931

Except in the Parallel searchi, Local search, and Open class, only a single core of one processor will be used for each entrant.

### Benchmark Selection

The benchmarks for MiniZinc Challenge 2016 (as well as for the qualification rounds) will be selected by the judges to try to examine some of the breadth of characteristics of FD solvers:

- propagation speed
- search speed
- global constraints
- satisfaction
- optimization

To obtain benchmarks of suitable difficulty we will select only such instances that can be solved by at least one of the participating solvers in a sensible time-frame. For the qualification rounds no such restriction applies.

In order to collect good benchmarks each entrant is strongly encouraged to submit one or two MiniZinc 2.0.13 models, making use of only the global constraints included in the MiniZinc 2.0.13 library, as well as some (preferably 20) instance files for each model. The instances should range from easy (about a minute) to hard (about 15 minutes) if possible.

Note that the model must conform to the problem format restrictions above.

Submitted benchmarks must be placed in the public domain.

### Initial Submission Round

There will be an initial submission round, which will provide the organizers with an opportunity to provide feedback on entries' compatibility with the competition hardware, compliance with the FlatZinc specification and any other matters that may arise. Submission in the initial round is not required in order to qualify for the final round, but it is **strongly** encouraged.

### The Challenge

The challenge will require solvers to process 100 MiniZinc models with a run-time limit of 20 minutes (process time) per problem.**NOTE** that the MiniZinc to FlatZinc/XML-FlatZinc time will be included in this time.

### Assessment

Each solver *s* is run on problem *p* and the following information is collected.

*timeUsed(p,s)*- the*wall clock*time used by the solver, or*timeLimit(p)*if it was still running at the timeLimit (quantized to seconds).*solved(p,s)*- true if a correct solution is returned, or correct unsatisfiability is detected*quality(p,s)*- the objective value of the best solution found by the solver (that is the last solution fully output before the time limit) assuming maximization*optimal(p,s)*- whether the objective value is proved optimal by the solver.

#### Scoring Procedure

The scoring procedure is based on the Borda count voting system. Each benchmark instance is treated like a voter who ranks the solvers. Each solver scores points equal to the number of solvers that they beat in the ranking (more or less). A solver*s*scores points on problem

*p*by comparing its performance with each other solver

*s'*on problem

*p*.

- If
*s*gives a better answer than*s'*it scores 1 point. - If
*s*and*s'*gives indistinguishable answers then scoring is based on execution time comparison (see below). - If
*s*gives a worse answer than*s'*it scores 0 point.

*s*and

*s'*,

*s*scores

*timeUsed(p,s') / (timeUsed(p,s') + timeUsed(p,s))*, 0.5 if both finished in 0s. The exception is that if

*solved(p,s)*is false, that is,

*s*fails to find any solution or prove unsatisfiability for problem

*p*it always scores 0 points (even if

*s'*also similarly fails).

#### SATISFACTION PROBLEM

A solver*s*answer is

**better**than solver

*s'*answer on satisfaction problem

*p*iff

*solved(p,s) && not solved(p,s')*

#### OPTIMIZATION PROBLEM

A solver*s*is

**better**than solver

*s'*on optimization problem

*p*iff

*solved(p,s) && not solved(p,s')*, or*optimal(p,s) && not optimal(p,s')*, or*quality(p,s) > quality(p,s')*, or

#### CLASSES

The scoring calculations will be done once for each run class: FD search, Free search, Parallel search, Open class, and Local search. Note that if too few solvers are entered in a class then the challenge won't be run for that class.

The organizers may well run entrants in the FD search class on a series of smaller problems to test that they indeed are following the given search strategy. These problems will not accrue points, but may disqualify an entry from the FD search class.

### Prizes

The solvers will be ranked on total points awarded. There will be prizes for the solvers with the highest scores in each of the run classes: FD search, Free search, Parallel search, Open class, and Local search. The organizers may also award prizes to the best solvers on a particular category of problems. Note that if too few solvers are entered in a class then the challenge won't be run for that class and no prizes will be awarded for that class.

### Restrictions

The organizers reserve the right to enter their own systems--or other systems of interest--to the competition, but these will not be eligible for prizes, but still will modify the scoring results. In addition, the organizers reserve the right not to run the challenge on classes with insufficient number of solver entrants.

Return to the MiniZinc Challenge 2016 home page.