هماهنگی فرایندهای توسعه محصول پراکنده شده : یک چشم انداز احتمالی از مدل سازی و طراحی پروژه

Managing worldwide supply pipeline operations concerns coordination and control of every step of the chain process starting from raw material sourcing, production, finally to distribution of market-specific items in retail places, all the way from product value inception and engineering, through manufacturing design, to the worldwide logistic planning. From systems perspectives, their relationships and interactions determine the overall performance. Coordinating such systems is very human-inclusive, characterised by abstract, ill-structured information interchange among well-partitioned expert groups. In this paper we documented the experience and implications of managing and modelling product development activities from a contingent perspective of interdependence. In our investigations amongst six international fashion corporations, crucial activity tasks in different countries were analyzed and evaluated within the context of launching schedule-driven fashion products. At the outset, we present the problem context, the issues arising from coordinating product development systems, and the approach we use to deal with the issues, i.e. modelling and manipulating the process interaction. We put forth a dependency-based process performance simulation, the related approach of data capture and the attribute constructs to represent the interactivity relationship. Finally, we discuss the computation process and evaluation strategy, which is indeed inspired by today's simulation-based optimization concepts. An effective GA heuristics is duly implemented.

This research examines the experience and implication of managing and modeling dispersed product development activities in the international fashion supply businesses. It encompasses an empirical investigation, which is based on the contingency perspective of organizational interdependence and restructuring; wherein the appropriate settings of bureaucracy and the respective interdependent process integration have to be established, subject to the kinds of uncertain innovation task and market environments (Clark and Fujimoto, 1991; Kusiak and Wang, 1993b; Levitt et al., 1999; Pennings, 1975; Victor and Blackburn, 1987). Pertaining to the contingency management needs for coordinating and restructuring organizational process, we posit activity task interdependence as a management contingency variable to model overall performance of a chained activity process. Interdependence is viewed as a vital part of systems that determines how organization activities co-operate and avail themselves to cope with changing market demand. Logically thought, some forms and patterns of activity interaction have to be preferred and perceived leading to better competitive and performance advantages. However, organization interactions are subtly intractable and carry types of intricate social and psychological attachments (Van de Ven and Walker, 1984; Uzzi, 1997). Notably, this contingency-viewed activity tasks are contextually dynamic and require teams to appreciate their distinctive objectives and constraints to proceed their activities in highly autonomous ways. Yet the action of theirs are interdependent, each relying upon an understanding of how the other teams are presumed to respond to the common environment concurrently. However, in reality, seldom single functional entity or enterprise is able to perceive all the perspectives throughout all phases of a business cycle, especially for today's marketplaces in which products are developed, sourced and distributed throughout worldwide supply pipelines (Abernathy et al., 1995; To et al., 2002). In this paper, the contingency-based research focuses particularly on a methodology that continually improves information flow process structure among complex, interdependent and geographically dispersed product development activities. In order to investigate the pertinent issues of integrating and managing such activities, the authors researched the activity workflow process structures in general and analyzed how the information was processed and managed among the activities in particular, in the context of global fashion businesses. To allow the generalization of empirical findings about activity workflow structure for product development, a flowchart model of activity processes has been established, as shown in Fig. 1. The flowchart corresponds with the apparel buying companies’ dynamic business planning processes and their nature of process and decision interdependency that arises in the course of developing global-oriented products. Generically, the product development involved several distinct plan work process systems: (1) anticipating fashion trends and product opportunities; (2) developing coordinated portfolio of product lines and their specifications; (3) budgeting and allocating fashion merchandise procurement; (4) organizing materials and logistic requirements for production; and (5) designing distribution and handling systems. In our observation the process systems are very human-tied and involved essentially 103 cross-country activity tasks. The cases studied represent today's virtual enterprising concepts; whereby fashion products are designed and developed collaboratively by globally dispersed functional teams and enterprises. All the operations along sorts of global supply pipelines are orchestrated and monitored in virtual form, and interdependently coordinated at the lowest levels of organization.

From the numerical results we obtained in the previous section, we are convinced that our framework is robust enough. However, the data needed in our workflow model are obtained from the development teams’ understanding and experiences, meaning that these data may not be usable on workflows that are vastly different from previous or existing ones. In other words, our methodology is more suitable for contingently reorganizing the coordination structure of existing development processes than for designing a radically new product development process (a similar problem was faced by Ahmadi et al. (2001), Smith and Eppinger (1997)). Hence, the modeling framework we present can form a learning mechanism by which companies can improve continually their activity planning and coordination. Currently, our methodology is justified by the results we performed on an observed number of data sets. Further justification can be achieved by more detailed computational experiments performed on randomly generated problems. However, the generation of random graphs with ‘realistic’ structures is a difficult problem and is a field of ongoing research. At this stage, the authors are still exploring ways to overcome this issue. Viewed from the angle of extendibility, the methodology and its outgrowing models are very flexible. For example, one can adopt an activity time model in which a design activity's duration may depend on other factors than the activity's number of invocations. See Carrascosa et al. (1998) for some inspiring ideas in this direction. One can also put additional restrictions on the choices of feedbacks in forming cut sets. A particularly useful restriction is to form cut sets with pairs of backward edges between pairs of interdependent tasks only. (Two activity tasks are interdependent if each of them is linked to the other by a feedback.) For each pair of feedbacks in a cut set, we may remove not both feedbacks, but removing one and turn the other into a forward edge. The authors address the following major issues in this research to achieve better performance of modeling globally dispersed activity processes, in terms of process design and control: Irregularity of information and data interchange among segregated functional and enterprise activity tasks makes the backward process flow likely and intractable. From an organization or functional structure perspective, the direction of information processing is not always in line with the activity task structural sequences that have conventionally been employed to model workflow processes and assign resources. Indeterminacy of the optimum process structural model stems from the progressively increasing evaluation complexity. Objective evaluation and verification of alternative activity process models cannot easily be made at the early stages of business activity planning. As such the performance of activities cannot be guaranteed or correctly reflected by the models, and does not take account of contingency management for process changes. This is the reason why the majority of activity planning modeling methods are advocated, but difficultly manipulated. Different types of activity interaction make process re-structuring difficult and complex. In most cases, removing or inhibiting some of feedback interaction among activities by policies entails the effective flow of information towards the process end goals. However, it is hard to select and justify where the feedback interaction should be treated (ignored), either by using process planner intuition or by using advanced computational tools. As mentioned beforehand, to define the respective product features and specification matching with the current supply availabilities, they should approach and negotiate with all possible domestic country importers and overseas manufacturers through the networks of affiliated overseas buying offices or trading agents. Then the proposed fabric materials would be developed, tested and evaluated at the major worldwide supply bases. Apparel accessories had to be acquired from other worldwide suppliers. Front end apparel manufacturers should spontaneously design the manufacturing process and machinery requirements for types of fabrics and accessories potentially to be selected. However, upon an anticipation of possible change in market demand or affordability, the retail store buyers or designers intuitively requested merchandising teams in various regional buying offices to revise the adoption of raw material, styling and sizing already agreed in the previous design processes. The manufacturing processes in respective supply countries correspondingly adjusted the design of production setup and re-scheduled raw material distribution and financial arrangement. Such changes would in turn affect the buyers to re-consolidate the end-market operations of promotion, distribution and pricing, in view of the consistency of global product, or store, image and quality offered to all marketplaces simultaneously. Inefficiency and ineffectiveness of activity communication and interactions at the stage of new product development would result in large magnitudes of change in the later stages of product manufacturing and distribution, so consuming significant resources and even leaving problems unresolved. Managing interactions among interdependent activities becomes the key issue for effective and timely new product design and development. The magnitude (strength) of an interactivity feedback that represents the importance of an interaction or information exchange among activities is difficult to measure. In essence, for pairs of activities, the vitality of feedback shows how beneficial or how important an earlier activity output is to the input of a following activity. This vitality is variable and changes with time and context; it is polysemous throughout stages of a process cycle. To measure and represent the magnitude in alternative process structure systems requires the assessment of psychometric values, which are themselves very judgmental and unstable from perceptual points of view. In recognition of the above, we attempt to integrate and mix some of the ideas discussed in the previous sections for managing and modeling product development activity workflow processes. These ideas are brought forward to develop a methodological framework in which cycle time of complex and dynamic activity process structures can be evaluated and improved in the time-efficient manner. While there were a few attempts (e.g. Ahmadi et al. 2001) to improve product development processes by changing the processes’ task structures (i.e. the execution sequences), we are not aware of previous studies specifically addressing the significance of change of information flow path and structure of activity tasks to project environments that require globally dispersed activity tasks to be completed interdependently in very short period, as illustrated in today's practice of mass customization and quick response in global textile marketplaces. In practical sense the methodology, especially in the aspects of simulation-based evaluation, is contemporary to study complex optimization problems (Fu 2002).