A generic coordination mechanism for lot-sizing in supply chains

Abstract

A new generic mechanism to coordinate decentral planning of a group of independent and self-interested decision makers, who are searching for an agreeable contract regarding multiple interdependent issues, in the case of asymmetric information is presented. The basic idea of the mechanism is that the group members cooperatively carry out an evolutionary search in the contract space. Therefore the (1, λ)-selection procedure, which is used in many evolutionary strategies, is combined with the Borda maximin voting rule, which has been applied successfully in group decision making. The proposed mechanism is realized, applied and evaluated for production coordination in a supply chain. A decentralized variant of the multilevel uncapacitated lot-sizing problem (MLULSP) is taken as the production model. For the evaluation 95 problem instances are generated based on MLULSP instances taken from the literature, with problem sizes varying from 5 to 500 items, from 12 to 52 periods. Experimental results show that the proposed mechanism is effective to determine fair cost distributions.

Coordination mechanisms with mathematical programming models for decentralized decision-making: a literature review

“…In other cases, with a finite solutions modelling approach (Distributed Multi-Level Uncapacitated Lot-Sizing Problem)(DMLULSP) they are voted and applied meta-heuristic neighbourhood search (Homberger et al, 2015) to find a better solution. These votes have also been evaluated by rules Borda maximum voting rule along with meta-heuristics ant colony or pheromone (Homberger et al, 2011(Homberger et al, , 2010 or evolutionary algorithms (Homberger, 2011), or with other voting rules Metropolis acceptance criterion and the meta-heuristic simulated annealing (Homberger, 2010) or with different rules Rawls or min-max voting rule and the meta-heuristic ant colony or pheromone (Buer et al, 2013). Also auctions using modelling approach of nonlinear mathematical programming (Tang et al, 2016b).…”

Integrating Side Payments into Collaborative Planning for the Distributed Multi-level Unconstrained Lot Sizing Problem

“…The distributed MLULSP (DMLULSP) was presented by [26] and assumes several local and selfish agents with private information instead of a single agent with full information. The existing collaborative solution approaches for the DMLULSP differ by their used metaheuristic search principles and the techniques used to aggregate the agents' preferences ( [26], [27] use simulated annealing, [28] uses an evolutionary strategy, and [3], [29], [30] use ant colony optimization). The DM-LULSP covers some important features of real world problems.…”

Collaborative machine scheduling: Challenges of individually optimizing behavior

“…Similarly, Fujita et al , use an approach that uses issue grouping and a limited amount of certain responses. Moreover, Homberger and Lang and Fink propose protocols that are inspired by evolutionary processes. Fink and Homberger develop a protocol that is based on an ant colony optimization algorithm.…”