Large-scale collaborative vehicle routing

Abstract Carriers can remarkably reduce transportation costs and emissions when they collaborate, for example through a platform. Such gains, however, have only been investigated relatively small problem instances with low numbers of carriers. We develop auction-based methods large-scale dynamic collaborative pickup delivery problems, combining techniques multi-agent systems combinatorial auctions. evaluate our approach in terms both solution quality possibilities strategic behaviour using real-world data set over 12,000 orders. Hence, this study is (to the best knowledge) first to assess benefits carrier cooperation propose an it. First, we use iterative single-order auctions investigate possible collaboration gains increasing Our results show that travel be reduced by up 77% 1000 carriers largely were previously observed smaller-scale collaboration. also ensure individual rationality guaranteed each auction. Next, compare multiple local established central auction mechanism observe proposed performs better on instances. Furthermore, improve quality, integrate two approaches allowing bundle system. analyze circumstances under which bundling beneficial decentralized system demonstrate cost 13% obtained Finally, whether vulnerable cheating: misrepresentation true values participants sometimes benefit them at collective. Although such not straightforward, discuss different means prevent