Meetings scheduling solver enhancement with local consistency reinforcement

Abstract

Meeting scheduling (MS) represents an important real-world group decision application that denotes one of the actual combinatorial problems. Solving this problem consists of scheduling all the meetings while satisfying all the constraints related to both the users and the meetings. However, given human nature, the solution is usually delineated by the encountering of conflicting preferences. Most of existing research efforts allow the relaxation of the users' preferences in order to reach an agreement between all the participants, which is not always possible. In addition, they do not deal with the achievement of any level of local consistency to enhance the efficiency of the solving process, and finally, they do not address the real difficulty of distributed systems, which is the complexity of message passing operations.Here we propose a new approach to facilitate and streamline the scheduling meetings process in any organization. This approach is based on the distributed reinforcement of arc consistency model, which takes into account the difficulties mentioned above. The present work focuses mainly on satisfying meetings hosts' preferences as much as possible, while taking into consideration all users' availability. The underlying selfish protocol is able to efficiently reach the best solution for the host of the meeting (according to the predefined criteria) whenever possible. This process is achieved with the minimal number of exchanged messages and while

A colored Petri net model for DisCSP algorithms

“…In Zivan and Meisels,44,45 comparisons between ABT, SBT, WCS, and some of their variants are done. One main result of these studies is consistent with the outcome of our paper; namely, the performance used in assessing DisCSP methods concern: the n-queens problem, [11][12][13]42 graph coloring, 11,12,43 meeting scheduling, 18,53 sensor networks, 36,37,53 and protein-folding problem. 27 In our study, the n-queens problem was utilized.…”

“…Moreover, authors use in their study a benchmark problem close to reality, namely, the distributed sensors network. Other problems frequently used in assessing DisCSP methods concern: the n‐queens problem, graph coloring, meeting scheduling, sensor networks, and protein‐folding problem . In our study, the n‐queens problem was utilized.…”

“…DisCSP proves to provide solutions for problems on resource allocation, [14][15][16] planning/scheduling and timetabling/meeting problems, [17][18][19][20][21][22][23][24][25][26] biochemistry, 27 multirobot planning and exploration, [28][29][30][31][32] and distributed configuration. 33,34 In the last years, an application area for DisCSP that had an important development regards various types of networks; one can find results on the use of DisCSP in cognitive radio networks, 35 wireless communication networks, 6,7 and sensor networks.…”

Privacy-Preserving Dialogues Between Agents: A Contract-Based Incentive Mechanism for Distributed Meeting Scheduling

“…Agents' preferences were quantified and different negotiation strategies were developed to improve the social welfare. By formulating the MAMS problem as a DCOP, the Max-Sum algorithms proposed in [8] and local consistency reinforcement [9] can be utilized where the solution space is traversed indirectly via limited private information exchange among all agents. In the work of [10], the extended VCG mechanism finds the optimal strategy of each agent, which is aligned with the solution of social welfare maximization.…”

Multi-variable distributed backtracking with sessions

“…-Timetabling/meeting problems [30,47,63]-Each participant wants to schedule several events at once. Each event is a variable, but constraints and other data for each participant must remain private and are thus encapsulated in an agent that represents a participant.…”