Real-world optimization problems often consist of several NP-hard combinatorial optimization problems that interact with each other. Such multi-component optimization problems are difficult to solve not only because of the contained hard optimization problems, but in particular, because of the interdependencies between the different components. Interdependence complicates a decision making by forcing each sub-problem to influence the quality and feasibility of solutions of the other sub-problems. This influence might be even stronger when one sub-problem changes the data used by another one through a solution construction process. Examples of multi-component problems are vehicle routing problems under loading constraints, the maximizing material utilization while respecting a production schedule, and the relocation of containers in a port while minimizing idle times of ships.
The goal of this competition in combination with its associated special session is to provide a platform for researchers in computational intelligence working on multi-component optimization problems. The main focus of this competition is on the combination of TSP and Knapsack problems. However, we plan to extend this competition format to more complex combinations of problems (that have typically been dealt with individually in the past decades) in the upcoming years.
The set of benchmarks used in this competition follows the idea of the "Travelling Thief Problem" (Mohammad Reza Bonyadi, Zbigniew Michalewicz, Luigi Barone: The travelling thief problem: The first step in the transition from theoretical problems to realistic problems.
IEEE Congress on Evolutionary Computation 2013: 1037-1044). Eucledian 2D Traveling Salesperson instances are combined with 0-1-Knapsack instances in such a way that it reflects aspects of problems from the real-world; for example, the total weight of the items in the knapsack influences the travel speed of a traveller. This introduced interdependence sets our benchmarks apart from capacitated vehicle routing problem instances, where this interdependence does not exist. For technical details on how the benchmarks were created, please see the manual (GECCO 2014 article)
A range of sample instances is available for researchers to experiment with before the final submission. The samples will include instances with few/many cities, with uncorrelated/correlated profits and weights, and instances with further characteristics.
In order to encourage researchers during the weeks before the submission deadline, we invite them to submit solutions for the sample instances. These results will be displayed online (without a reference to the authors) and then will serve as performance indications for other researchers.
These files will be run on our state of the art Linux servers. The evaluation criteria are
final solution quality after a fixed budget of fitness function evaluations (to mimic "expensive" evaluations), and (cancelled)
- final solution quality after a fixed budget of computation time (to mimic the importance of "deadlines").
Currently, we are aiming at a maximum runtime limit of 10 minutes per problem.
Additional technical details will be published online in time.
The Australian-based company Optimatics
is providing the total amount of AUD 1,000
Optimatics provides the water industry's most powerful water planning decision support tools and processes to enable optimal water services.
The company enables decision makers to optimize the planning, operations and management of their water and wastewater networks with Optimatics' proprietary software, consulting and distributed computing solutions.
While we do not require this, we strongly encourage the participants to register for WCCI 2014 due to the advantages.
- Your algorithm needs to be executable from the command line and it should only take the instance file as a parameter, e.g.,
java thisIsMyAlgorithm instances/a280_n279_bounded-strongly-corr_02.ttp
- You can assume to have at least 16 GB of RAM at your disposal.
- We will stop you program after 10 minutes. This means that your algorithm needs to write the best result to a file before
the 10 minutes are over.
Note that you can of course try to write out the results frequently, but since we will potentially execute your code in parallel (not mixed with other competitors!), it is in your own interest that you do not overload the file system with very frequent writes.
- Output: the file with the result for a single iteration should follow the following naming convention: <ttpfile>.<algorithmname>.<systemtime>, e.g.,
- The file needs to contain the following information as comma-separated values in square brackets: the permutation of the cities in the first line (note: the first city is "1", do not print the "1" at the end), and the list of packed items in the second line (the numbering of the items starts with "1"). For example:
which means that only the items with the numbers 20 and 113 are picked, and the sequence of visited cities is <1,5,4,2,3,1>. Note that this format can easily be achieved in Java with the function Arrays.toString(...).
Please submit your code via email
Your submission needs to include a README.txt that will help us to run your code. For example, it needs to contain an example of how we can execute the code on the command line of our Linux machines.
We have Matlab, Java, and Mono installed in case you are using our Matlab/Java/C# code. If you intend to submit code in a different language, please check with us first.
We can help just little with debugging, so if you have tested your code thoroughly, chances are that we will enjoy the organisational part more.
February 2014: instances and code online
16th June 2014, 11:59pm (Hawaii-ean time): submission deadline for algorithms
Please contact us if you want to get notifications via email.
9th July 2014: we have announced the winners!
1. Prize (AUD 500) Dmitriy Artemkin
2. Prize (AUD 250) Yi Mei
2. Prize (AUD 250) Kok Wooi Hew
We cannot publish here details about the approaches used by the participants, as they intend to publish their approaches.
What the best approaches have in common is the following. In the first step, the approaches construct a TSP tour via 2-OPT or Lin-Kernighan, and in the second step items are picked based on "maximise the profit" or on a "first fit" basis (with a focus on the last cities in the tour).