EXHIBIT 99.264


[LOGO]                         MINUTES OF MEETING

                  Topic: Discussion of CONG Open Issues


                  DATE:   4/22/97
                  TIME:   10:00 - 4:00 P.M.
                  PLACE:  MESA PAVILLION HILTON, MESA, ARIZONA

================================================================================

ATTENDEES:         Ziad Alaywan, Farrokh Albuyeh, Ongun Alsac, George Angelidis,
                   Reynaldo Bernal, Vladimir Brandwajn, Joe Bright, Paul Gribik,
                   Murray Nixon, Mauro Prais, Brian Stott, Au Vojdani

PREPARED BY:    Farrokh Albuyeh
APPROVED BY:    Mi Vojdani


                                     TOPICS

LOAD DISTRIBUTION AND CONGESTION PARAMETERS:

o        There will be a single set of default prices for load curtailment. If a
         load That is bid does not include The cost for curtailment, the default
         set would be used. Similarly for generators, the curtailment prices are
         used when it is found necessary to reduce the generator below its
         minimum capacity. Typically, the variable parts of price curves (for
         load & generator) are bid. For extension beyond The high and low limits
         the default values are used.

                                     [GRAPH]

o        The issue of load distribution factors was discussed. It was agreed
         That only one set of distribution factors (day type dependent) will be
         use for all portfolios to distribute the load of the zone or group into
         individual loads. ABB to write up the resolution and distribute to SA
         User Group for review by 4/25. The User Group to review by 4/30.

o        It was agreed that loads for a given SC in a given zone, or load group
         would be treated as blocks during optimization, if their definition in
         the portfolio is at the zone or load group level.

o        Agreed on how to handle infeasible situations. The procedure requested
         by the customer is such that the inter-zonal MW flow constraints should
         be considered hard constraints at any point during the optimization
         process. If all resources have been exhausted and the inter-zonal MW
         violations persist, the unit MW limits should be relaxed.



o        ISO/SA Users Group confirmed that all incremental/decremental cost
         curves are staircase curves. One single staircase curve is used for
         increments and for decrements.

o        It was agreed that the same default curves shall be used whenever the
         incremental/decremental curve provided in a portfolio does not cover
         the complete range specified by the unit MW limits or when the unit MW
         limits has to be relaxed due to infeasibiity.

o        It was agreed that the default curves will be part of the Congestion
         Manager parameters, and will not come as part of any Schedules.

REDISTRIBUTION CONSTRAINTS:

o        The issue of Redistribution Constraints will be addressed as follows.
         For all units of a scheduling coordinator in a zone, compute an
         equivalent cost curve. Use this cost curve to make sure that we have a
         monotonically increasing incremental cost curve so that the program
         would not attempt to redispatch generators of a scheduling coordinator.
         Paul Gribik to write specific details on how this curve clearance price
         normalization is performed.

o        The handling of identical units (i.e., similar cost characteristics)
         was discussed. PCA's standard method for handling of identical units is
         adequate.

PERFORMANCE:

o        PCA expressed caution that if most of the portfolios are rich (i.e.,
         many generators selling to many loads), then market separation
         constraints may create a performance problem. PCA is designing a
         methodology to overctnne this problem Additionally, the fact that in
         most of the cases load components in a portfolio will be given at a
         zone or at a group will reduce the performance problem drastically.
         Therefore, this is not a major concern at this point. (After the
         meeting, Paul Gribik demonstrated that the LP bases will stay sparse
         with an easily treated structure. This observation should address PCA's
         concern regarding market separation constraints giving rise to a dense
         basis matrices in the LP problem.)

INTER-ZONAL AND INTRA-ZONAL CONGESTION MANAGEMENT INTERACTIONS:

o        Compatibility between Inter-Zonal and Intra-Zonal system representation
         is a concern. It was agreed that this is a Stage II issue, and that
         changes in the design of Intra-Zonal CONG in Stage II will not affect
         the fbnctionality and the design of the Inter-Zonal CONG in Stage I.

o        Interactions between Inter-Zonal and Intra-Zonal will not be
         automatic it will be handled through manual intervention. The
         Intra-Zonal CONG will provide sensitivity of external generation to the
         interface flows to guide this manual process.

o        It was agreed that Intra-Zonal CONG will be solved for the entire
         system, and not on a zone by zone basis.

ANCILLARY SERVICES INTERFACE:

o        It was agreed that neither of the two alternatives that were
         communicated to the ISO Alliance as substitutes for the existing
         congestion management protocol in DSOW can be implemented by 1/1/98.
         Therefore the issue of the interface between Ancillary Service
         Management and CONG was dropped.

OTHER ISSUES:

o        Section 9.7, pages 81- If schedules are similar and are to be
         curtailed, they are curtailed pro-rata.






                                     ISSUES:

None

                                  ACTION ITEMS:

o        ABB to outline their understanding of how bus loads for each Scheduling
         Coordinator is going to be calculated starting from the Load Forecast,
         and submit to SA Users Group for review by 4/25. The SA User Group to
         review the outline and comment by 4/30.

         ABB to issue a change regarding OS

o        ISO to provide inputs by 4/25 regarding true up of losses

o        ABB to propose a method to calculate losses.





NEXT MEETING:

4/30/97, PG&E, 77 Beale, Room 1582