The Mixed Multi-Unit Combinatorial Auctions Test Suite
Supply Chain Formation (SCF) is the process of determining the participants in a supply chain, who will exchange what with whom, and the terms of the exchanges. Mixed multi-unit combinatorial auctions (MMUCAs) offer a high potential to solve SCF problems, and thus be employed for the automated assembly of supply chains of agents. MMUCATS is a test suite for MMUCAs that allows researchers to test, compare, and improve their winner determination algorithms for MMUCAs.
Auction visualization
MMUCATS provides several graphical facilities for the structural analysis of WDP instances. Thus, it allows to depict: (i) the supply chain structure along with the distribution of goods and transformations between tiers; (ii) the bid graph structure capturing the relationships among bids, goods, stock goods, and goods required as a result of the supply chain operation; (iii) the transformation dependency graph showing the dependencies among transformations; and (iv) the strongly connected components of the transformation dependency graph.
Bid generation
MMUCATS encloses an algorithm to generate artificial data sets that are representative of the sort of scenarios a WD algorithm is likely to encounter. The algorithm takes inspiration on the structure of supply chains. Interestingly, the flexibility of our generator makes possible that within the very same supply chain we can find varying degrees of production complexity, different production structures involving goods from multiple tiers, and varying degrees of competitiveness. Although MMUCATS encloses an implementation in MATLAB of the generator, it allows users to incorporate their own generators as long as the generator's output is an XML file that complies with the MMUCATS DTD for WDP instances.
Supply chain formation
MMUCATS allows users to incorporate their implementations to solve the WDP instances created by the artificial data set generator. Thus, external calls to solvers using either GLPK or CPLEX (solver implementations included) are straightforward from the MMUCATS user interface. MMUCATS interprets the solutions output by either GLPK or CPLEX to graphically display the optimal structure of the supply chain, the net benefit of the formation process, the time employed by the solver, and the number of decision variables employed. Once solved an instance of the WDP, MMUCATS can animate the solution to show users how transformations are used along the optimal supply chain and how goods are produced and consumed. This feature is aimed at helping users how the resulting supply chain is expected to behave when enacted.
Supply chain generation
MMUCATS allows to automatically generate the workflow for the optimal supply chain structure produced by a winner determination algorithm as an ISLANDER specification. Two major benefits stem from this facility. Firstly, the ISLANDER specification can be readily employed to enact a supply chain as an electronic institution where winning bidders participate with the aid of the Electronic Institutions Development Environment (EIDE). Secondly, users can also benefit from the simulation facilities provided by EIDE to analyse the expected performance of the resulting supply chain.
Download instructions
Click here to download the latest version of MMUCATS.
To run MMUCATS simply do:
- Unzip the downloaded file.
- Go to the lib directory.
- Double click the mmucap-viewer.jar file or run "java -jar mmucap-viewer.jar" from your command line.
Java 1.6 is required to correctly execute the application.