The Defense Department's effort to design the communications network for a space-based missile defense, widely considered the paramount technical challenge of the Strategic Defense Initiative research program, has been hobbled by mismanagement and inefficient spending for two years, according to a new report by the General Accounting Office.

The GAO said after a year-long study that the Strategic Defense Initiative Organization (SDIO) "needs to improve its ability to provide timely and effective management direction and oversight" for the effort, aimed at providing the equipment needed to control hundreds of space weapons and sensors in a defense against Soviet intercontinental ballistic missiles.

"The problems, if not corrected, could reduce SDIO's ability to provide needed information for an informed . . . decision" in the early 1990s on whether a space-based missile defense should be deployed, the GAO said. The Reagan administration has set the early 1990s as the target for such a decision.

The GAO report, released late last month in response to a request by House Appropriations defense subcommittee Chairman Bill Chappell (D-Fla.), points up fresh concerns about the scientific underpinning of the "Star Wars" program, an initiative that will be part of talks this week between President Reagan and Soviet leader Mikhail Gorbachev.

In December 1985, a panel of expert scientific advisers, noting the inherent difficulty of building equipment to manage a battle between U.S. and Soviet space weapons, missiles and nuclear warheads, called the task "the paramount strategic defense problem."

The GAO report observed that "this particular part of the SDI effort is frequently criticized as being unattainable." Critics have argued that the computer and communications requirements for split-second attacks on thousands of potential Soviet targets in space are simply beyond the capability of foreseeable U.S. technology.

Experts agree that without precise coordination, an attack by U.S. defensive weapons against swiftly moving Soviet warheads and decoys would probably fail.

SDI officials have spent $605 million -- or about one-tenth of their total budget -- on "battle management" research since 1984, according to the GAO.

But many research contracts awarded to private industry have been canceled before work was complete because of sudden shifts in program priorities or decisions by SDI officials to siphon battle management funds to other programs, the report said.

An SDI spokesman acknowledged that "there were some inefficiencies {in battle management research} caused by program realignment," but attributed the problems to budget constraints imposed by Congress and said improvements had been made since GAO auditors finished their work earlier this

year.

The GAO, an arm of Congress, said senior Navy, Air Force and SDI research managers "pointed to SDIO's inadequate direction and planning" as the major cause of slow progress in the program, despite official Pentagon claims that staff shortages and inadequate funding were to blame.

The GAO said, for example, that outside experts had urged the agency to design the overall missile defense system with battle management needs in mind, but said "SDIO made little progress in accomplishing {this} . . . for nearly two years."

Instead, SDI officials developed overall designs for a missile defense system in space with little concern for potential shortfalls in battle management equipment, the GAO said. Once this oversight was recognized, some work had to be redone at a cost of more than $16 million.

The GAO said SDI officials decided early this year to delay "indefinitely" a final design for the missile defense system "because of the need to better ensure . . . integration" of battle management equipment in the design. The Army, which has coordinated some key battle management research, "may have to rework its ongoing . . . effort" after spending more than $32 million, the GAO said.

The report also cited complaints from research officials that progress was slowed because of "poor program documentation and limited dissemination of . . . information." It said SDI officials never drafted a formal plan to protect against "fraud, waste or mismanagement."

The report said, for example, that SDI officials spent $1 million to develop experimental computer software and then decided to develop computers that could not use it. In another instance, they failed to stop expenditures of $5 million on computer-related research that was ordered by mistake.

SDI officials attempted last April to improve battle management research, but "the extent to which this will be successful remains to be seen," the GAO said. It will remain unclear at least until the latter half of next year, the agency added.