How OR helped pacify Bosnia


OR's contribution to the NATO-led IFOR responsible for implementing the internationally agreed peace accord in Bosnia-Herzegovina

by Nigel Cummings

ORS has recently received a collection of declassified NATO documents concerning operation Firm Endeavour and its effect during the Bosnia and Herzegovina conflicts which began in 1995. Headquarters Allied Command Europe Rapid Reaction Corps (HQ ARRC) were deployed at these locations as the land component of the NATO led Implementation Force (IFOR). Its task was to oversee the military aspects of the internationally agreed peace accord.

By the time IFOR troops has started to arrive in theatre (December 95) a cease-fire had been in place for some months. The situation remained fragile however with factional armies still in place across the wartime confrontation lines. In general there was little freedom of civilian movement at this time, and some of the population depended entirely on international aid for staple foods and medicines.

Included as an integral part of the HQ was a small Operational Analysis Branch (OAB). The OAB was deployed on 22nd December 1995, and stayed in theatre until November 20th 1996. The OAB was a small unit of five 3 civilian analysts, 1 military officer and a non-commissioned officer. The initial emphasis of IFOR was those crucial tasks which greatly reduce the likelihood of the resumption of fighting between former warring factions and IFOR. These tasks were essentially completed within 5 months of the implementation of the OAB. Although many military tasks continued throughout the full year long deployment, there was a realisation that overall IFOR success depended on greater emphasis on civil support.

Black military helicopter flying low over grasslands.

Throughout the year the OAB was augmented with short term augmentees (six week periods) from the UK, Canada and the NATO Consultation Command and Control Agency (NC3A). The OAB was seen as being at the top of an analytical and scientific iceberg, and able to call on additional support and expertise from the UK and NATO NC3A.

Throughout the planning time the HQ had undergone a series of exercises to assess command structures, designs of HQs etc. OAB had taken part in these activities. At one point OAB software and expertise were called upon to provide early analysis as justification for troop numbers required to fulfil the mission.

It must have been quite a shock for the civilian component of the OAB to find themselves coping with first aid, nuclear, biological and chemical drills and weapon handling. Status was never really made clear. N. J. Lambert, OA Branch, HQ ARRC, BFPO 40 explains: We were a small group of civilians in a military environment. There was even some surprise when we said we would deploy! Our status was never completely clear; for example, should we wear civilian or combat style clothes, what were our equivalent military ranks? In fact we had always intended to wear military clothes without badges or rank but with Operational Analysis epaulettes. Certainly this was a very sensible decision bearing in mind where we ended up, and the security situation.

The OABs mission was to support the Land Component Commander of IFOR (Commander ARRC): To give independent analytical and scientific advice to the Commander to aid his decision-making over the spectrum of ARRC activities.

In essence the main tasks conducted by the OAB whilst on Operation Firm Endeavour were to enable the commander to have access to valuable background information and data to which he would not normally have access information he could use to counter speculation that nothing was changing or getting worse.

The OAB also assisted logistics staff with the task of transition planning the various options for withdrawal and or redeployment of IFOR. HQ staff were also assisted with the difficult task of information management, a task which encompassed collection and collation of data, storage and retrieval of data, analysis and methods of displaying such data. The OAB also had the responsibility of collecting and archiving data relating to the campaign. However when data was collected which was perceived as being of value to the greater historical analysis OA community, it was collected (albeit on an ad-hoc basis).

Data collected by the OAB was categorised as follows.

  • Data collected to support key OA studies.
  • Data collected as by-product of OAB information management assistance.
  • Data collected through archiving activities.
  • Data sought out by OAB for future OA studies.

The OAB used 7 different methods of data collection to achieve their aims. Surveys were resorted to when it became obvious that no other data sources existed in theatre to fulfil the data requirement for a task. In many cases data required to address a particular OA task was already present, so it was merely a case of scoring existing data sources.

Where specific data was required to calibrate or define data for a specific mathematical model, Trials and Observations were the methods employed. For example, the data used to calibrate a system dynamics model of deployment used observations of the roulement of units in June 1996.

Another method used was that of creating Specific Task databases. The hidden agenda behind the OABs helpfulness with HQ bases in the creation of such databases, was not just to assist in the management of critical information but also to collect operational data in magnetic format and within a database.

Archiving was another method in use during the OABs operational life in Bosnia-Herzegovina. Discrete archiving, where a snapshot could be collected of a valuable data set was used alongside continuous archiving methods where the data changed on a periodic basis. Indirect archiving was also utilised on a few occasions where data required continuous periodic archiving. In this case the OAB would have to rely on an individual to forward the data direct for archiving or to archive the data in-house and await a visit by OAB personnel. Procurement of data was another method in use.

Some of the saddest losses of data were due to ignorance as to the importance of data. In one case archived data on actual convoy timings was lost following a handover between two officers despite an agreement to continue the data collection. In other cases data was shredded or deleted by staff who had forgotten or who were ignorant of the importance of the data.

The IT equipment employed was quite sophisticated for the time - the OAB made use of 3 standard Pentium 90 PCs with 32Mb of RAM networked with Microsoft Windows for Work Groups. Two CD writers were also employed for data storage and the network server was a Sun Sparc 10, with a spare held for backup. Two Pentium 90 based laptops were also used as analytical PCs or mobile presentation units. Regrettably the IT equipment was prone to frequent failure, which resulted in a great deal of OAB resources being addressed to the maintenance and upgrade of failing computers.

In early February 1996, the question of the mechanics of IFOR for redeployment was addressed with the formation of the Theatre Transition Planning Group (TTPG). This was formed to effect a seamless transition between IFOR and its follow-on force. In this role the TTPG had executive power to task organisations as required. As a matter of course the OAB was invited to provide input to the development of this plan and was formally tasked to provide a mechanism to assess the robustness of proposed force redeployment plans for the IFOR.

Soldier with a camouflage helmet looking past the camera, other soldiers are behind in the background.

This was seen as a classic Operational Research question with the construction, calibration and use of a computer model being the most appropriate approach. Initially only a single member of the branch was addressed to the task. Preliminary work highlighted the need for a model which could be used in support of redeployment planning to have the following characteristics.

  • The model would have to be quick running.
  • The model would have to be transparent to the user.
  • The model would have to produce output that could be readily understood by the non technically minded.

Although a number of approaches were available, System Dynamics (SD) formed the basis for the construction of the model. In essence this method is the mathematics of flows and their influence on quantities varying between particular states. SD models have attracted some criticism because their ability to look at high level problems is often ill defined, so much so that it may be impossible describe a process completely in terms of rates and flows. This was not felt to be a concern by the OAB though, as the problem was seen to be comparatively low level and had already been well defined.

The package chosen for the model was the Ventana Simulation Environment Vensim which the branch had taken with them as an integral part of their toolkit. A discussion with the planning staff indicated that any model which was to be used should have the following characteristics.

  • A representation of each of the distinct categories of equipment to be redeployed.
  • A representation of the process of preparing equipment to commence redeployment.
  • A representation of the constraints on the redeployment process caused by finite numbers of movement assets.
  • A representation of the constraints on the redeployment process caused by finite route capacity and staging area capacity available for the redeployment.
  • A representation of the embarkation process and the associated constraints imposed by port capacity.
  • A representation of the effect of IFOR redeployment timelines of a follow on force deploying into the theatre whilst IFOR was redeploying.
  • A representation of the effect of varying weather conditions, movement asset availability, convoy density/speed, driver day length and varying route occupancy levels.

In the event a model was eventually produced which proved to be credible with military staff, to support the decision making process and to provide a means of problem definition and clarification. The technicalities involved in the production of the model are too immersive to cover within the scope of this article, though the declassified NATO reports I have been using throughout, will soon be available for perusal from the OR archive at the Modern Records Centre, Warwick University Library, Coventry.

The OAB found that during the operation, the overhead associated with the gathering of data was considerable a task in its own right. Long term military operations generate huge amounts of data that require collection, collation and manipulation. The OAB felt that with hindsight a data collection plan might have helped with the gathering of data. It was also noted that the lack of a central database made the task of accessing the data required for analysis much more difficult than it needed to be.

Considering the size of the OAB unit, it is a testament to the efficiency of OR staff manning the unit that so many tasks were carried out. The OAB unit had to measure former warring factions and the general framework agreement for peace. Methodology here was based on database and GIS analysis with results expected within hours. Traffic surveys utilising statistical analysis of raw and scored data were also employed, transition planning utilising critical path analysis and fast running system dynamics models had to be up and running in the minimum amount of time, with the initial model being produced within one month.

The OAB unit was also responsible for much of the information management required, and its staff were on hand to offer scientific advice and analysis too especially where problems were perceived to be messy and/or complex, or entailing considerable uncertainty, in which case the pragmatic solutions offered by OR disciplines in the OAB unit were much in demand.

First published to members of the Operational Research Society in Inside O.R. June 1999