Abstract

The shifting paradigm of supply chain management is manifesting increasing reliance on automated collaborative planning and event monitoring through information-bounded interaction across organizations. An end-to-end support for the course of actions is turning vital in faster incident response and proactive decision making. Many current platforms exhibit limitations to handle supply chain planning and monitoring in decentralized setting where participants may divide their responsibilities and share computational load of the solution generation. In this thesis, we investigate modeling and solution generation techniques for shared commodity delivery planning and event monitoring problems in a collaborative setting. In particular, we first elaborate a new model of Multi-Depot Vehicle Routing Problem (MDVRP) to jointly serve customer demands using multiple vehicles followed by a heuristic technique to search near-optimal solutions for such problem instances. Secondly, we propose two distributed mechanisms, namely: Passive Learning and Active Negotiation, to find near-optimal MDVRP solutions while executing the heuristic algorithm at the participant's side. Thirdly, we illustrate a collaboration mechanism to cost-effectively deploy execution monitors over supply chain network in order to collect in-field plan execution data. Finally, we describe a distributed approach to collaboratively monitor associations among recent events from an incoming stream of plan execution data. Experimental results over known datasets demonstrate the efficiency of the approaches to handle medium and large problem instances. The work has also produced considerable knowledge on the collaborative transportation planning and execution event monitoring.