Cem M. Baydar

Automated Generation of Robust Error Recovery Logic in Assembly Systems Using Genetic Programming

Abstract Automated assembly lines are subject to unexpected failures, which can cause costly shutdowns. Generally, the recovery process is done “on-line” by human experts or automated error recovery logic controllers embedded in the system. However, these controller codes are programmed based on anticipated error scenarios and, due to the geometrical features of the assembly lines, there may be error cases that belong to the same anticipated type but are present in different positions, each requiring a different way to recover. Therefore, robustness must be assured in the sense of having a common recovery algorithm for similar cases during the recovery sequence. The proposed approach is based on three-dimensional geometric modeling of the assembly line coupled with the genetic programming and multi-level optimization techniques to generate robust error recovery logic in an “off-line” manner. The approach uses genetic programmings flexibility to generate recovery plans in the robot language itself. An assembly line is modeled and from the given error cases an optimum way of error recovery is investigated using multi-level optimization in a “generate and test” fashion. The obtained results showed that with the improved convergence gained by using multi-level optimization, the infrastructure is capable of finding robust error recovery algorithms. It is expected that this approach will require less time for the generation of robust error recovery logic.

Prediction and Diagnosis of Propagated Errors in Assembly Systems Using Virtual Factories

Large-scale automated assembly systems are widely used in automotive, aerospace and consumer electronics industries to obtain high quality products in less time. However, one disadvantage of these automated systems is that they are composed of too many working parameters. Since it is not possible to monitor all these parameters during the assembly process, an undetected error may propagate and result in a more critical detected error. In this paper, a unique way of detecting and diagnosing these types of failures by using Virtual Factories is discussed. A Virtual Factory was developed by building and linking several software modules to predict and diagnose propagated errors. A multi-station assembly system was modeled and a previously discussed ‘‘off-line prediction and recovery’’ method was applied. The obtained results showed that this method is capable of predicting propagated errors, which are too complex to solve for a human expert. @DOI: 10.1115/1.1411966#

One-to-one modeling and simulation: a new approach in customer relationship management for grocery retail

The ever-increasing competition in retail industry puts pressure on retailers to deal with their customers more efficiently. Currently most companies use Customer Relationship Management (CRM) systems to maximize the customer satisfaction level by trying to understand more about their behaviors. However, one disadvantage of the current approaches is that they focus on the segmentation of customers into homogenous groups and they disregard examining the one-to-one relationship of each individuals behavior toward each product. Therefore, individual behavior cannot be captured in detail. Modeling individual behavior for each product enables several strategies of pricing by keeping the customer satisfaction at the maximum level. One example is offering a personal discount on a particular item to a customer who is price sensitive to that particular product. Therefore, you can still sell other products at the non-discounted level to this customer by keeping him satisfied. In this paper, individual pricing approach is discussed. The aim of this study is to develop a conceptual framework to analyze the feasibility of individual pricing. Customer behaviors can be modeled individually with respect to each product for a grocery store. Several factors can be used to determine these behaviors such as customers need, brand loyalty and price sensitivity. Each customer can be modeled as an adaptive agent using qualitative descriptions of behaviors (i.e., highly price sensitive). Then, the overall shopping behavior can be simulated using a multi-agent Monte-Carlo simulation. It is expected that with this approach, retailers will be able to determine better strategies to obtain more profits, better sales and better customer satisfaction.