Industrial & Systems Engineering
Decomposition of Large Scale Man-power Planning Problems: Assignment, Loading and Scheduling
Model decompositions for manpower planning problems utilizing time-based hierarchical structuring of the decisions and predictor-corrector iterative control schemes are investigated. Our proposed method consists of 3 stages: (1) determination of the number of employees for stochastic demand data in a given service region, (2) the loading problem in which jobs are distributed to appropriately skilled employees, (3) scheduling to shifts, subject to work rules and back-up requirements.
Network Simplex-like Algorithms for Stochastic Appointment Scheduling
Scheduling appointment times when appointment durations are random is a common problem. Given the sequence of appointments, and scenarios generated from duration distributions, optimizing appointment starting times is an LP. We develop a “network simplex-like” algorithm to solve it, and contrast with previous algorithms based on the L-shaped method.
Stochastic Sequencing & Scheduling of Operating Rooms
We study sequencing and scheduling problems for a single OR when surgery times are random variables. We first present a network flow based algorithm that solves the scheduling problem for a given sequence of surgeries with multiple scenarios. Next we address the sequencing problem using different heuristic methods.
Risk-averse Policies in a One-Warehouse Multiple-Retailer System with Demand and Supply Uncertainty
We examine demand and supply uncertainty in a One-Warehouse Multiple-Retailer system and compare centralized and decentralized inventory strategies. We show that while risk pooling generates lower expected costs in a centralized strategy, risk diversification generates lower risk in a decentralized strategy. We demonstrate that under certain risk-averse objectives, decentralizing inventory is the optimal policy for the OWMR system.
Supervisory Control of a Multi-Echelon Supply Chain: Structure, Modeling, Performance Measures and System Analysis for Inter-organizational Control
Petri nets are frequently utilized to model system dynamics due to their ability to handle concurrencies and sequential dependence. In this paper a portion of the Supply Chain Operations Reference (SCOR) model has been extracted and modeled using Petri nets for the purpose of exerting supervisory control upon a multi-echelon supply chain. The activities of source, make and deliver, inherent in the SCOR model form the basis of the representation of the Petri Net model for each echelon considered in the supply chain model. A supervisor is placed above the base model of each echelon to exert local constraints. These constraints are at the tactical and operational levels. An Enterprise level SCM is added which enforces additional constraints consisting of long term planning goals at the strategic level. Invariant analysis is used to create the supervisors. Performance measures of the supply chain as one entity are formulated to determine the effectiveness of any partnership. An efficient method for finding the current state of the system is developed which is used to determine the performance measures of each echelon. This paper presents an approach to the overall structure and Petri Net modeling for the system and is intended to extend the use of supervisory control from a shop-floor level to an inter-organizational facility and enterprise level.